Syncsort MFX

Programmer’s Guide

Version 3.1
Syncsort MFX Programmer’s Guide

© 2010, 2020 Syncsort Incorporated. All rights reserved.

This document contains unpublished, confidential, and proprietary information of Syncsort

Incorporated. No disclosure or use of any portion of the contents of this document may be made
without the express written consent of Syncsort Incorporated.

Version 3.1 - Version 8

Last Update: July 20, 2020
 Table of Contents

Summary of Changes .............................................................................................. i

Precisely Rebranding........................................................................................... i
Summary of Changes for Syncsort MFX Release 3.1 ........................................ i

Chapter 1: Introduction ....................................................................................... 1-1

An Introduction to Syncsort MFX .................................................................. 1-1
MFX’s Basic Functions .................................................................................... 1-1
MFX’s Data Utility and SortWriter Features ............................................... 1-3
MFX Operational Features ............................................................................. 1-6
Syncsort MFX Value-Added Products ............................................................ 1-7
Structure of the Programmer’s Guide ............................................................ 1-7
Related Reading............................................................................................... 1-9
Online Message Help..................................................................................... 1-10

Chapter 2: MFX Control Statements .............................................................. 2-1

Control Statement Summary Chart ............................................................... 2-3
Disk Sort, MAXSORT, and PARASORT Control Statement
Requirements ................................................................................................ 2-11
Data Utility Processing Sequence................................................................. 2-12
Control Statement Examples ........................................................................ 2-14
Rules for Control Statements ....................................................................... 2-14
ALTSEQ Control Statement ........................................................................ 2-19
DUPKEYS Control Statement ..................................................................... 2-21
END Control Statement ............................................................................... 2-28
INCLUDE/OMIT Control Statement ........................................................... 2-29
INREC Control Statement ........................................................................... 2-54
JOIN Control Statement ............................................................................... 2-57
JOINKEYS Control Statement ..................................................................... 2-59
MERGE Control Statement ......................................................................... 2-66
MODS Control Statement ............................................................................ 2-80
OMIT Control Statement ............................................................................. 2-84
OUTFIL Control Statement ......................................................................... 2-85
OUTREC Control Statement ..................................................................... 2-134
RECORD Control Statement ..................................................................... 2-227
REFORMAT Control Statement ................................................................. 2-231
SORT Control Statement ............................................................................ 2-234
SUM Control Statement ............................................................................. 2-254

Syncsort MFX Programmer’s Guide i

Chapter 3: How to Use the MFX Data Utility Features ............................. 3-1
Sample Data Utility Applications................................................................... 3-1
Selecting Input Records .................................................................................. 3-3
Selecting Relevant Fields from the Input Records ........................................ 3-6
Combining Records within a File ................................................................. 3-11
Joining Records from Multiple Files............................................................. 3-14
Making Output Records Printable and Easy to Read ................................. 3-25
Dividing a Report into Sections ................................................................... 3-41
Writing Headers and Trailers for a Report ................................................. 3-43
Totaling and Subtotaling Data .................................................................... 3-51
Obtaining Maximum, Minimum and Average Data ................................... 3-57
Counting Data Records ................................................................................. 3-59
Creating Multiple Output Files ................................................................... 3-63
E-mailing a Report in PDF Format .............................................................. 3-66

Chapter 4: JCL and Sample JCL/Control Statement Streams ................. 4-1

EXEC Statement ............................................................................................. 4-3
For MAXSORT, PARASORT, DB2 Query and MULTIIN Support .............. 4-3
Coding Conventions for DD Statements ........................................................ 4-3
STEPLIB/JOBLIB DD Statement ................................................................. 4-4
SYSOUT DD Statement ................................................................................. 4-5
SORTIN DD Statement .................................................................................. 4-5
SORTINnn or SORTINn DD Statement ....................................................... 4-7
SORTJNF1 and SORTJNF2 DD Statements................................................. 4-9
SORTOUT, SORTOFxx, SORTOFx, SORTXSUM, and
SORTXDUP DD Statements .......................................................................... 4-9
SORTWKxx or SORTWKx DD Statement .................................................. 4-11
SYSIN DD Statement ................................................................................... 4-12
$ORTPARM DD Statement .......................................................................... 4-12
SORTCKPT DD Statement .......................................................................... 4-15
For Exit Routines that Require Link-editing at Execution Time .............. 4-15
DD Statements for MAXSORT, PARASORT, DB2 Query and
MULTIIN Support ......................................................................................... 4-17
Sample JCL/Control Statement Streams .................................................... 4-17

Chapter 5: PARM Options ............................................................................... 5-1

Precedence Rules ............................................................................................ 5-1
PARM Option Summary Chart ...................................................................... 5-2
Additional PARMs .......................................................................................... 5-4
MFX PARM Options ........................................................................................ 5-5

Chapter 6: Invoking MFX from a Program ................................................... 6-1

Programming Flexibility vs. Performance ..................................................... 6-1
DD Statements ................................................................................................ 6-2
Invoking the Sort/Merge from an Assembler Program ................................. 6-2
The 24-Bit Parameter List .............................................................................. 6-4

ii Syncsort MFX Programmer’s Guide

 Sample Assembler Invocation Using 24-Bit Parameter List ...................... 6-11
The 31-Bit Extended Parameter List .......................................................... 6-12
Sample Assembler Invocation Using 31-Bit Parameter List ...................... 6-17
The 64-Bit Parameter List ...................................................................... 6-18
Syncsort Java Class ....................................................................................... 6-22

Chapter 7: The Coding and Use of Exit Programs ...................................... 7-1

What Is an Exit? .............................................................................................. 7-1
Loading the Exit Routines into Main Storage ............................................... 7-3
Exit Conventions ......................................................................................... 7-3
Register Conventions ...................................................................................... 7-4
The Exit Communication Area ...................................................................... 7-5
Exits E11, E21, and E31 - Preparing for Other Exit Routines .................... 7-6
Exit E32 - Invoked Merge Only: Creating Input Records ............................ 7-6
Exit E14 - Deleting, Summing, Changing Records ....................................... 7-7
Exit E15 - Creating, Revising, or Analyzing the Input File ......................... 7-8
Coding a COBOL E15 Exit Routine.............................................................. 7-11
Example 1: Fixed-Length Records ............................................................... 7-12
Example 2: Variable-Length Records .......................................................... 7-13
Coding a C E15 Exit Routine ........................................................................ 7-20
Fixed-Length Records - Function Definition ................................................ 7-20
Variable-Length Records - Function Definition ........................................... 7-22
Exit E25 - Deleting, Changing, and Summing Records .............................. 7-28
Exit E35 - Adding, Deleting, and Changing Records .................................. 7-29
Coding a COBOL E35 Exit Routine ............................................................. 7-32
Coding a C E35 Exit Routine ....................................................................... 7-43
Fixed-Length Records - Function Definition ................................................ 7-43
Variable-Length Records - Function Definition ........................................... 7-45
Exit E16-Taking Action on Insufficient Intermediate Storage .................. 7-51
Exits E17, E27, and E37 - Closing Data Sets .............................................. 7-52
Exits E18, E38, and E39 - Checking Labels, Processing Read or
Write Errors, End-of-File Routines, Special VSAM Processing ................. 7-52
Exit E61 - Modifying the Collating Process ................................................ 7-57
Coding REXX Exits ....................................................................................... 7-59

Chapter 8: The Flow of the Sort ..................................................................... 8-1

Chapter 9: MAXSORT........................................................................................ 9-1

MAXSORT: A Maximum Capacity Sort ........................................................ 9-1
MAXSORT’s Advantages................................................................................. 9-4
Job Control Language ..................................................................................... 9-4
DD Statements ................................................................................................ 9-5
SORTBKPT DD Statement ............................................................................ 9-6
SORTOU00 DD Statement ............................................................................. 9-7
SORTOUnn DD Statements ........................................................................... 9-8
Using Disk for Intermediate Output .............................................................. 9-9

Syncsort MFX Programmer’s Guide iii

 SORTCKPT DD Statement .......................................................................... 9-10
Control Statements ....................................................................................... 9-10
PARM Options .............................................................................................. 9-11
Exit Programs ............................................................................................... 9-14
Invoking MAXSORT from a Program .......................................................... 9-15
Restarting MAXSORT ................................................................................... 9-15
MAXSORT’s Operator Interface .................................................................. 9-16
Sample MAXSORT JCL/Control Streams ................................................... 9-18

Chapter 10: PARASORT .................................................................................. 10-1

PARASORT: Parallel Input Processing for Elapsed Time
Improvement ................................................................................................. 10-1
PARASORT Applicability.............................................................................. 10-1
Job Control Language ................................................................................... 10-2
DD Statements .............................................................................................. 10-3
SORTIN DD Statement with PARASORT .................................................. 10-4
SORTPARn DD Statements ......................................................................... 10-6
Special Channel Separated Esoteric Names ............................................... 10-7
Sortwork Considerations ............................................................................... 10-9
Operations Notes ........................................................................................... 10-9

Chapter 11: MFX DB2 Query Support ............................................................ 11-1

Restrictions .................................................................................................... 11-1
Job Control Language.................................................................................... 11-2
DD Statements............................................................................................... 11-2
SORTDBIN DD Statement ........................................................................... 11-3
PARM Options .............................................................................................. 11-4
Operation........................................................................................................ 11-4
Record Description......................................................................................... 11-5
Record Description: Trial Mode Execution................................................... 11-6
Sample MFX DB2 Query Application........................................................... 11-8

Chapter 12: Multiple Input Files ...................................................................... 12-1

Restrictions .................................................................................................... 12-2
Job Control Language.................................................................................... 12-2
Operation........................................................................................................ 12-4
Sample Multiple Input File JCL and Control Statements .......................... 12-5

Chapter 13: The Dictionary Feature............................................................... 13-1

Building a Dictionary .................................................................................... 13-2
The Constant_name Statement: Rules and Syntax ..................................... 13-5
The Field_name Statement: Rules and Syntax............................................ 13-9
The Operator Statement: Rules and Syntax .............................................. 13-15
Activating the Dictionary Feature .............................................................. 13-19
Using Dictionary_names in MFX Control Statements .............................. 13-20
Specifying Dictionary Listing Output......................................................... 13-24

iv Syncsort MFX Programmer’s Guide

 Error Handling for Dictionary Statements ................................................ 13-24
Sample Dictionary Statements ................................................................... 13-25
Using JCL SET and PROC Symbols to Create
Dictionary_Names ....................................................................................... 13-29

Chapter 14: Performance Considerations..................................................... 14-1

Disk Sort? MAXSORT? PARASORT? .......................................................... 14-1
JCL Sorts vs. Program-Invoked Sorts ......................................................... 14-2
Control Statement Issues ............................................................................. 14-2
The Efficient Use of PARMs ......................................................................... 14-3
Optimizing System Resources....................................................................... 14-3
The Incore Sort ............................................................................................. 14-4
Disk Space Considerations............................................................................ 14-4
The Coding and Use of Checkpoint-Restart ................................................ 14-7
Automatic Checkpoint-Restart ..................................................................... 14-8
Deferred Checkpoint-Restart ........................................................................ 14-9
Optimizing Data Set Placement ................................................................ 14-10

Chapter 15: The HISTOGRM Utility Program ............................................. 15-1

What Is HISTOGRM? ................................................................................... 15-1
Using HISTOGRM to Determine L6 and L7 Values for MFX .................... 15-1
Control Parameters for HISTOGRM ........................................................... 15-2
Job Control Language ................................................................................... 15-5
Executing HISTOGRM through an E15 Exit .............................................. 15-5
HISTOGRM Messages ................................................................................ 15-10

Chapter 16: Value-Added Products ............................................................... 16-1

Syncsort PROCSort – An Accelerator for SAS® Sorting ............................ 16-1
Syncsort PipeSort ......................................................................................... 16-2
Syncsort ZPSaver .......................................................................................... 16-2

Chapter 17: Messages ....................................................................................... 17-1

Syncsort MFX Messages................................................................................ 17-1
Syncsort ZPCopy Messages ......................................................................... 17-71
Syncsort ZPCompress Messages ................................................................. 17-71
Syncsort ZPSort Messages .......................................................................... 17-72
Syncsort PROCSort Messages .............................................................. 17-72
License Key Messages ................................................................................ 17-73

Chapter 18: Diagnostics and Technical Support ....................................... 18-1

Troubleshooting Abends ................................................................................ 18-1
Before Contacting Precisely Support ............................................................ 18-5
Contacting Precisely Support........................................................................ 18-6

Index ............................................................................................................... Index-1

Syncsort MFX Programmer’s Guide v

vi Syncsort MFX Programmer’s Guide
 Summary of Changes

This chapter summarizes the changes for Release 3.1 of Syncsort MFX.

Precisely Rebranding
Syncsort products have been rebranded to Precisely brand names. Syncsort
MFX was formerly known as MFX for z/OS.

Summary of Changes for Syncsort MFX Release 3.1

This section contains information that is included in the Syncsort MFX
Programmer’s Guide and Syncsort MFX Installation Guide for Release 3.1,
which supports IBM z/OS Version 2 Release 3 and earlier.

Syncsort Java Wrapper Class

The Syncsort() Java class provides a wrapper around the Syncsort MFX sort/
merge/copy utility. When invoked from Java, the input data can be taken from
either existing data sets or produced directly from a Java application.
Likewise, the output data from Syncsort MFX can be written to a data set or
read directly back into the Java application.
For more information, refer to "Chapter 6, Invoking MFX from a Program”.

Syncsort MFX Programmer’s Guide i

SORTWORK Encryption
During conventional SORT processing, all data written to SORTWORK data
sets is not encrypted. The new SWCRYPT parameter enables encryption for
SORTWORK data sets to meet modern information security and compliance
standards. The SORTWORK encryption feature can be configured during
installation as a SYNCMAC option and passed as a run-time parameter.
For more information about SWCRYPT as a SYNCMAC option, refer to the
Syncsort MFX Installation Guide.
For more information about SWCRYPT as a run-time parameter, refer to
"Chapter 5, PARM Options”.

Block Support for E15 and E35 User Exits

The E15 and E35 user exits in the 64-bit parameter list have been enhanced to
support the transfer of blocks of records between MFX and the E15/E35 exits.
For more information, refer to “The 64-Bit Parameter List” on page 6-18.

Performance Improvements
The following functions will see improvements in both CPU and elapsed
times:
• Sort applications with a simple OUTREC statement that reduces the
output record length.
• Certain large JOIN applications.

PARM Options
Run-time Parameters
• OPTELAP – This new option enables or disables running offload work on
the general CP to improve the elapse time when all zIIP engines are busy.
• RESERVEZ – Reserves a specified amount of memory above the 2-
gigabyte bar for the user.
• SWCRYPT – This new option enables or disables the use of encryption for
SORTWORK data sets.
• ZIIPPCT – This new option allows you to define the zIIP offload
percentage within a range of 1-100 percent.
For more information, refer to "Chapter 5, PARM Options”.

ii Syncsort MFX Programmer’s Guide

Messages
The following new message explanations are in "Chapter 17, Messages”.

Message Description
WER292A BLOCK E15/E35 ERROR, REASON CODE nn

EXPLANATION: A Block Exit error was found in the E15 or E35 routine
module, where nn represents the error reason code.

The Block Exit error reason codes are:

• 01 Block List Parameter Area or Block List Address Invalid
• 02 Block List Type Invalid
• 03 Block List Boundaries Invalid
• 04 Block Address (Type 2) or Record Address (Type 1) Invalid
• 05 Block Exit Return Code Invalid
• 06 Record Length Invalid (Variable-length Record)

WER559I UNSUPPORTED FUNCTIONS WITH SORTWORK ENCRYPTION

EXPLANATION: SORT is not using encryption mode for SORTWORK

because of unsupported functions.

WER560I SORTWORK ENCRYPTION USED

EXPLANATION: SORT is using encryption mode for SORTWORK. All

data in the SORTWORK data set is encrypted.

WER561A IBM CALLABLE SERVICE XXXXXXXX ERROR, RETURN CODE

XXXXXXXX, REASON CODE XXXXXXXX

EXPLANATION: A SORTWORK encryption process error occurred

during an IBM call statement.

ACTION: For more information about the IBM return and reason codes,
refer to the IBM publication z/OS Cryptographic Services Integrated
Cryptographic Service Facility Application Programmer's Guide, SC14-
7508-06. Or contact Precisely Support for assistance.

Syncsort MFX Programmer’s Guide iii

iv Syncsort MFX Programmer’s Guide
Chapter 1 Introduction

An Introduction to Syncsort MFX

Syncsort MFX is a high performance sort/merge/copy utility. It is designed for the
advanced facilities of the zSeries architecture and exploits the features of the z/OS
operating system, but also supports the system architectures of IBM System/390
and compatible computers.
MFX is designed to conserve system resources, provide significant performance
benefits, and operate efficiently in 31-bit or 64-bit environments.
MFX can be initiated through job control language or invoked from a program
written in COBOL, PL/1, or Assembler language. A JCL-initiated sort is more
efficient because MFX totally controls the sort execution, including I/O
management and main storage management. Exit routines may be written in
COBOL, C, FORTRAN, REXX, or Assembler language to give a JCL sort additional
programming flexibility. Exits may also be in PL/1 when MFX is invoked by a PL/1
program.

MFX’s Basic Functions

MFX has three basic functions:
• Sorting – rearranging data set records to produce a specific sequence.
• Merging –combining up to 100 pre-sequenced data sets into one data set which
has the same sequence.
• Copying –reproducing a data set without going through the sorting process.

Syncsort MFX Programmer’s Guide 1–1

Sorting
A sort rearranges the records in a data set to produce a specific sequence, e.g.,
chronological or alphabetic order. MFX provides the following sorting techniques:
• Disk Sort, the standard sorting technique. Information in the Programmer’s
Guide refers to the Disk Sort unless otherwise indicated.
• MAXSORT, a maximum capacity sorting technique with an enhanced
breakpoint/restart capability. MAXSORT can sort any collection of data -
regardless of size - using a limited amount of disk space. MAXSORT is
described in the MAXSORT chapter of this guide.
• PARASORT, a sorting technique that significantly reduces elapsed time for
sorts whose input is a multi-volume tape data set and/or concatenated tape
data sets. PARASORT improves performance by using multiple tape drives in
parallel. PARASORT is described in the PARASORT chapter of this guide.
A sort logically consists of four phases that perform the following functions:
• The control statements and JCL information are read and analyzed and the
operational parameters for the sort are established.
• The input data is read into main storage and sorted.
• If necessary, intermediate results are written to temporary storage devices.
• The sorting process completes and the sorted data is written to the specified
output device(s).

Merging
A merge combines up to 100 pre-sequenced data sets into one data set which has
the same sequence. A merge has two phases that perform these functions:
• The control statements and JCL information are read and analyzed and the
operational parameters for the merge are established.
• The files are merged and the merged data is written to the specified output
device(s).

Copying
A copy reproduces a file, completely bypassing the sorting process. A copy has two
phases that perform these functions:
• The control statements and JCL information are read and analyzed and the
operational parameters for the copy are established.
• The copied file is written to the specified output device(s).

1–2 Syncsort MFX Programmer’s Guide

MFX’s Data Utility and SortWriter Features
MFX is designed to improve programmer productivity by reducing the time the
programmer/analyst must spend designing, testing, and debugging applications.
With MFX’s extensive Data Utility and SortWriter features, data processing
applications previously requiring several steps can be accomplished in a single
execution.
MFX’s Data Utility features include a join facility, a multiple output facility, a full
range of report writing capabilities, and a facility to create a PDF, RTF or HTML
output file and e-mail it to one or more recipients. There are also many record
selection and record reformatting facilities. These options allow you to design sort/
merge/copy applications that can accomplish a host of related tasks.

Join Processing
The join facility, controlled by the JOINKEYS, JOIN, and REFORMAT control
statements, joins records from two source files. When you join the records from two
files, each record from the first file (the left side) with a given value in a specified
field (the join key) is joined to each record from the second file (the right side) with
the identical value in a specified field in that record. Thus, if m records from the
left side have a given join key value, and n from the right side have the same join
key value, the join results in m*n records with that join key value.
Options are provided to control several aspects of the join operation.
• Specification of the placement of the data fields within the record created by the
join operation is provided through the REFORMAT control statement. This
allows you to specify which fields from the two records are to be placed in the
joined record. Partially or completely missing fields will be filled into the
resulting joined record by the use of a specified pad byte.
• Record selection via the INCLUDE/OMIT parameter of the JOINKEYS
statement eliminates records from either or both of the two input files prior to
join processing.
• Specification of whether the join input data is already sorted per the
JOINKEYS control fields is controlled by the SORTED parameter on the
JOINKEYS control statement. If the join input data set is already sequenced
according to the specified JOINKEYS fields, the overall performance of the
application improves.
• Inner join, left outer join, right outer join, and full outer join are all supported
through the use of the JOIN control statement.

Syncsort MFX Programmer’s Guide 1–3

Generating Multiple Output
The multiple output facility (OUTFIL) allows multiple output files to be generated
with just one pass of the sort. Each of these files can have unique specifications
that determine which records are to be included, how the records are to be
formatted, and which report capabilities are to be used. Moreover, all these files
can be written to the same output device, or each can be written to a different
device.

Creating Reports
MFX’s SortWriter feature (OUTFIL) allows you to design comprehensive reports
easily and efficiently. SortWriter options allow output data to be flexibly formatted
with headers and trailers, which can include data fields. Various kinds of numeric
results can be produced at report, page, and section levels. These include totals,
subtotals, minimums, subminimums, maximums, submaximums, averages,
subaverages, record counts, and subcounts. Output record fields can be realigned;
the records can be padded with blanks, characters, and binary zeros; and numeric
data can be converted and edited. Automatic pagination, page numbering, and
dating are also provided.

Creating PDF, RTF, and HTML Files and E-mailing Them

The OUTFIL OUTPUT feature can be used to create an output file in PDF, RTF or
HTML format. Any of these files can be e-mailed as an attachment to one or more
recipients.

Selecting Records, Reformatting Records, and Summing Fields

Record selection, record reformatting, and summing are other important MFX Data
Utility features. Record selection via the INCLUDE/OMIT feature permits certain
records to be included or omitted from an input data set based on comparisons
between two data fields or between a data field and a constant. Date data formats
work with the CENTWIN option to ensure that century evaluation is applied to
INCLUDE/OMIT comparisons involving 2-digit year data.
Record reformatting after input and/or before output, provided by the INREC/
OUTREC capability, allows you to delete or repeat portions of records; insert
spaces, characters, binary zeros, date constants and sequence numbers; realign
fields; convert numeric data to its printable format; and convert data to its
printable hexadecimal format. The CENTWIN option and date data formats enable
conversion of 2-digit year fields to printable or packed decimal 4-digit years of the
appropriate century. The ability to delete irrelevant fields before sorting via INREC
can provide important performance benefits. Additionally, a variable-length record
format input file can be converted into a fixed-length format output file or a fixed-
length record format input file can be converted into a variable-length format
output file.

1–4 Syncsort MFX Programmer’s Guide

 The SUM feature allows records with equal sort control fields to be deleted and
optionally sums numeric fields on those records. The deleted records can optionally
be written to a separate data set.
The DUPKEYS feature provides all the facilities of the SUM feature plus the
ability to replace specified numeric fields in the retained record with the minimum,
maximum, or the average value of the field for all records with the same control
field.

Sample SortWriter Report

The report below illustrates the versatility of MFX’s Data Utility and SortWriter
features. First, irrelevant records are omitted from the input file and the input
record is reformatted to eliminate unnecessary data fields. Then the file is sorted
by invoice status and invoice date. The output record is reformatted for readability
and the numeric fields are converted and edited. The report itself is divided into
sections and subsections based on control field breaks. Headers and trailers
identify the data fields, provide record counts and section and cumulative totals,
and include the date and page number.

PAGE: 3
ACCOUNTS RECEIVABLE AGING REPORT FOR 01/30/92 DATE:04/22/92
*********************************************

INVOICE STATUS:O
*****************

------------ ------- ---- ----- -------- --------------------- ---------------------

INVOICE BALANCE
COMPANY NAME ADDRESS CO # INV # INV DATE PRODUCT TAX PRODUCT TAX
------------ ------- ---- ----- -------- --------------------- ---------------------
REPUBLIC DATA NYC NY 2681 86013306 1/17/91 1,100.00 90.75 1,100.00 90.75
RICE FEATURES CHI IL 2244 86013298 1/17/91 1,500.00 75.00 1,500.00 75.00
SIDNEY COLLEGE HOU TX 4762 86013297 1/17/91 2,500.00 150.00 2,500.00 150.00
WINIFRED INDUST WAS DC 1177 86013299 1/17/91 650.00 26.00 650.00 26.00
PIZZUTO LOANS STL MO 4633 86022200 2/15/91 550.00 22.00 550.00 22.00
RICE FEATURES CHI IL 2244 86022198 2/15/91 1,500.00 75.00 1,500.00 75.00
SIDNEY COLLEGE HOU TX 4762 86022197 2/15/91 500.00 30.00 500.00 30.00
REGENCY TRUST CO BOS MA 4986 85124011 12/15/91 1,500.00 75.00 1,500.00 75.00
SIDNEY COLLEGE HOU TX 4762 85124016 12/15/91 5,000.00 300.00 5,000.00 300.00

---------- ---------- ---------- ----------

TOTAL NUMBER OF INVOICES: 11 MONTHLY TOTALS: $22,850.00 $1,484.50 $22,850.00 $1,484.50
---------- ---------- ---------- ----------

BALTIC AVENUE CORP CLE OH 0636 86022207 2/15/91 650.00 29.25 650.00 29.25
FASTEROOT EQUIP BAL MD 4980 86022205 2/15/91 1,700.00 76.50 1,700.00 76.50
FEDERAL FABRICS SHV LA 5143 86022204 2/15/91 1,750.00 70.00 1,750.00 70.00
PATIO PRODUCTS MRY CA 3029 86022203 2/15/91 850.00 51.00 850.00 51.00
TURENIUS FOR. EXCH. DTT MI 8325 86022201 2/15/91 1,600.00 64.00 1,600.00 64.00
WINES ASSOCIATES SMF CA 1794 86022209 2/15/91 750.00 45.00 750.00 45.00
DESIGN TECHNOLOGIES LAX CA 2520 85124017 12/15/91 360.00 21.60 360.00 21.60
POLL DATA CORP LAX CA 0846 85124019 12/15/91 600.00 36.00 600.00 36.00

---------- ---------- ---------- ----------

TOTAL NUMBER OF INVOICES: 8 MONTHLY TOTALS: $8,260.00 $393.35 $8,260.00 $393.35
---------- ---------- ---------- ----------

Syncsort MFX Programmer’s Guide 1–5

DB2 Query Support
MFX can directly retrieve data from a DB2 database based upon a user provided
query. An SQL SELECT statement is used to specify the criteria of the request, and
the query of the DB2 database will be in place of MFX's SORTIN or E15 processing.
SORT or COPY, but not MERGE, functions can be used with DB2 queries. All MFX
features that are performed after E15 processing are available for use with the
DB2 query facility.
This feature improves performance over DB2’s DSNTIAUL program by allowing
DB2 data to be passed directly into a SORT or COPY operation, without the use of
setup steps or the need for user-written E15 exits. Refer to “Chapter 11, MFX DB2
Query Support” for more information.

MFX Operational Features

MFX will take advantage of data space and memory objects to further improve
performance. A portion of the address space may be allocated for the MFX ZSPACE
technique. This technique was created as a replacement for hiperspace. It allows
native use of the available central storage resources. This technique eliminates the
additional overhead produced when hiperspace is simulated by the z/OS operating
system in a z/Architecture environment. It provides superior CPU performance and
reduced system overhead compared to a conventional hiperspace application.
MFX can also interact with exits and invoking programs such as VS COBOL II,
COBOL/370, C370 V2R1 with V2R2 C370 library, SAA AD/Cycle C370 Release 2,
and IBM C/C++ V3R2 programs.
MFX’s PARMEXIT feature permits the dynamic modification of PARM values
based on the conditions at execution time. This feature facilitates the passing of
additional parameters to specific jobs.
Other operational features include resident, reentrant code, interactive and
streamlined installation and maintenance procedures; automatic release or
secondary allocation of direct access intermediate storage (SORTWK) and output
(SORTOUT) space without JCL specification; dynamic allocation of SORTWK
space under z/OS (DYNALLOC); and automatic incore sorting.

1–6 Syncsort MFX Programmer’s Guide

Syncsort MFX Value-Added Products
Value-added products available from Syncsort MFX can significantly improve
performance:
Syncsort PROCSort - An Accelerator for SAS® Sorting is a high performance,
transparent replacement for the SAS procedure PROC SORT. Compared to PROC
SORT, Syncsort PROCSort reduces the resources required for sorting within SAS
applications and cuts sort elapsed time.
Syncsort PipeSort enables MFX to run multiple sorts simultaneously on the
same input data. For large input files, Syncsort PipeSort significantly reduces total
elapsed time compared to running separate sort jobs.
Syncsort ZPSaver is a set of enhanced technologies for MFX to offload copy, SMS
compression and sort processing to zIIP processors, effectively reducing the
workload on the main CPU. Specifically, Syncsort ZPSaver includes:
• Syncsort ZPCopy is a high-performance, transparent replacement for copy
processing initiated with SORT FIELDS=COPY (including BetterGener).
Syncsort ZPCopy is designed to execute almost all copy jobs on zIIP engines and
utilize parallel processing for I/O with Parallel Access Volume (PAV) technology.
• Syncsort ZPCompress offloads to zIIP engines the CPU time that is
associated with SMS Version 1 compression when reading non-VSAM SORTIN
and/or writing non-VSAM non-OUTFIL SORTOUT data sets with SMS
compression attributes.
• Syncsort ZPSort is a high-performance, transparent replacement for sort
processing that executes almost all sort steps on zIIP engines.
• Syncsort ZPEncrypt enables the use of SORTWORK encryption in the Sort in
zIIP application for offloading CPU time to zIIP engines.
For more detailed information regarding each of these products, see “Chapter 16
Value-Added Products”.

Structure of the Programmer’s Guide

The Syncsort MFX Programmer’s Guide is a reference manual designed for
applications programmers who are using Syncsort MFX to sort, merge, or copy
sequential data sets. This manual is self-contained and assumes only a basic
working knowledge of the operating system and its job control language. It should
not be necessary to refer to any other manual to produce an efficient sort.

Syncsort MFX Programmer’s Guide 1–7

 MFX Control Statements describes how to specify and use the ALTSEQ,
DUPKEYS, END, INCLUDE/OMIT, INREC/OUTREC, JOIN, JOINKEYS, MODS,
OUTFIL, RECORD, REFORMAT, SORT/MERGE, and SUM statements. The
discussion of a particular control statement includes these topics: the statement’s
syntax format, the versatility provided by the various parameters (many of which
are unique to MFX), and the interaction between the control statement and other
statements.
How to Use MFX’s Data Utility Features explains and illustrates the Data
Utility and SortWriter features through a series of sample applications. Each
application is self-contained and provides instructions for specifying both the
required JCL and the appropriate control statements.
JCL and Sample JCL/Control Statement Streams analyzes MFX’s job control
requirements and describes the MFX DD statements, each of which is illustrated
with an example. JCL and control statement streams for MAXSORT and
PARASORT are also described. Numerous examples are provided.
PARM Options describes the operational parameters of MFX and identifies the
delivered defaults. This chapter explains how to specify such features as dynamic
allocation of SORTWK space under z/OS, automatic secondary allocation and
release of SORTWK space, the ability to skip a certain number of records or stop
after sorting a certain number of records, and message routing.
Invoking MFX from a Program describes MFX invocation through assembler
programs using 24-bit, 31-bit, and 64-bit parameter lists. Numerous examples are
provided.
The Coding and Use of Exit Programs indicates at which points during sort
processing user-written exit routines can be executed. Each exit point is fully
documented together with the appropriate tasks. Examples of COBOL E15 and
E35 exit routines for fixed and variable-length records are included.
The Flow of the Sort provides a skeletal view of the flow of control in the standard
Disk Sort (including the incore sort), merge and copy. This chapter indicates the
order in which the control statements and exit routines are processed, information
which is particularly useful at the design stage of an application.
MAXSORT explains when MAXSORT should be used, describes its JCL
requirements, control statements and PARM options, and provide examples. The
chapter also examines MAXSORT’s restart capability and its operator interface.
PARASORT explains the elapsed time advantages of the technique, the type of
applications where it can be applied, and the JCL requirements.
MFX DB2 Query Support explains how MFX can improve performance by
allowing DB2 data to be passed directly into a SORT or COPY operation without
the use of setup steps or user-written E15 exits.
Multiple Input Files explains how to specify multiple VSAM and non-VSAM data
sets as input to MFX with the MULTIIN facility. The MULTIIN facility can be used
for a SORT or COPY application.

1–8 Syncsort MFX Programmer’s Guide

 The Dictionary Feature describes how to create symbolic dictionary names for
fields, constants and output columns, and use those dictionary names in MFX
control statements.
Performance Considerations describes how to design the most efficient
application. It contrasts the merits of Disk Sort, PARASORT, and MAXSORT, JCL
and invoked sorts, the incore sort, and standard SORTWK techniques. Formulas
for calculating main storage and SORTWK requirements are provided. Other
topics include the efficient use of control statements and PARMs, tuning main
storage and SORTWK allocations and the use of the Checkpoint-Restart feature.
The HISTOGRM Utility Program describes how to use the HISTOGRM program
to report on the composition of variable-length files. This program indicates the
average record length, byte total, record total, block count and record count. Job
control requirements, control statements and messages are outlined. Sample job
streams illustrate how to run HISTOGRM as a separate job and as an E15 exit
during a variable-length sort.
Value-Added Products describes Syncsort PROCSort – An Accelerator for SAS®
Sorting, Syncsort PipeSort, and Syncsort ZPSaver technologies. This chapter also
provides detailed information regarding their functions and special features.
Messages documents all of the WERnnnx messages generated by the Syncsort
MFX program.
Diagnostics and Technical Support includes sections describing
“Troubleshooting Abends” and “Before Calling Precisely Support”.

Related Reading
The following guides supplement the information provided in the Syncsort MFX
Programmer’s Guide.
Syncsort MFX Installation Guide
This manual explains how to install and maintain Syncsort MFX and defines the
default options.
Exploiting MFX: SortWriter Data Utilities Guide
This two-part user’s guide demonstrates how MFX’s versatile Data Utility features
provide an efficient, one-step alternative to writing, testing and debugging
programs. Five comprehensive sample applications illustrate how the control
statements work together to produce formatted reports.

Syncsort MFX Programmer’s Guide 1–9

 Exploiting MFX: MAXSORT
This user’s guide explains how to use the special MAXSORT feature of MFX to sort
very large amounts of data with only a limited amount of disk space. MAXSORT’s
unique restart capability is described and sample job control streams and tuning
information are included.
Exploiting MFX: JOIN
This booklet demonstrates through several examples various approaches for
creating applications with the join facility. Each example contains a statement of
the problem, a sample of the inputs, MFX control statements used to produce the
output, and a sample of the output.

Online Message Help

All Syncsort MFX messages and their explanations can be accessed online through
an ISPF/PDF dialog. Contact your system administrator for information about the
operation of the message help facility.

Online Help

1–10 Syncsort MFX Programmer’s Guide

Chapter 2 MFX Control Statements

The control statements tell MFX how to process files. There are sixteen control
statements:

Control
Statement Function

ALTSEQ Specifies an alternate collating sequence for control fields with an AQ

(pg. 2-19) format.

DUPKEYS Deletes records with equal SORT or MERGE fields; optionally calculates
(pg. 2-21) the sum, minimum, maximum, or average values of numeric fields with
equal SORT or MERGE fields.

END Signals the end of control statements.

(pg. 2-28)

INCLUDE Specifies the criteria which determine whether or not records are
(pg. 2-29) included in an application.

INREC Reformats the input record before sort/merge processing.

(pg. 2-54)

JOIN Specifies the disposition of paired and unpaired records in a join.

(pg. 2-57)

JOINKEYS Enables join feature processing and identifies the fields used to select
(pg. 2-59) records for join processing.

Table 1. MFX Control Statements

Syncsort MFX Programmer’s Guide 2–1

 Control
Statement Function

MERGE Defines a merge or copy application and specifies merge control fields.
(pg. 2-66)

MODS Specifies user exit(s).

(pg. 2-80)

OMIT Specifies the criteria which determine whether or not records are omitted
(pg. 2-84) from an application.

OUTFIL Describes the output file(s) and specifies SortWriter and processing
(pg. 2-85) options.

OUTREC Reformats the output record after sort/merge processing.

(pg. 2-134)

RECORD Provides record information at various processing stages.

(pg. 2-227)

REFORMAT Defines the record layout to be produced by join processing.

(2-231)

SORT Defines a sort or copy application and specifies sort control fields.
(pg. 2-234)

SUM Deletes records with equal SORT or MERGE fields and sums numeric
(pg. 2-254) fields on those records.

Table 1. MFX Control Statements (Continued)

2–2 Syncsort MFX Programmer’s Guide

Control Statement Summary Chart
The following table summarizes the parameters of each control statement and
indicates default values.

Control
Statement Parameters Delivered Default
Name

ALTSEQ Standard EBCDIC series

CODE=(ccpp 1 [,ccpp 2 ]...)

DUPKEYS No calculation of fields;

  no reduction of equal-
 function [ ,function ]... [ ,FORMAT=f ] 
 FIELDS=NONE  keyed records
 ALLDUPS 
  [ ,XDUP ]
 FIRSTDUP [,NODUPS] 
 LASTDUP [,NODUPS] 
 NODUPS 
 

where function is:

 AVG 
 MAX 
  = ( p 1 ,l 1 [ ,f 1 ] [ ,p 2 ,l 2 [ ,f 2 ] ]... )
 MIN 
 SUM 

XDUP

END

INCLUDE Sort/Merge all records

 ALL 
 NONE 
 
  ,AND,  
COND=   ,&,  
 (c   c 2 ... ) [,FORMAT=f] 
 1 
 ,OR, 
  ,|,  
 

INREC Input records unchanged

   
  FIELDS  
 [ PARSE=(subparm), ]  BUILD =(fields) 
  OVERLAY  
   
 IFTHEN=(subparm) [,IFTHEN=(subparm), ... ] [ ,IFOUTLEN=n ] 
 FINDREP=(subparm) 

Table 2. (Page 1 of 8)Control Statement Summary Chart

Syncsort MFX Programmer’s Guide 2–3

Control
Statement Parameters Delivered Default
Name

JOIN F1

F2

ONLY

UNPAIRED Join only paired records

JOINKEYS
FIELDS =(p 1 ,l 1 [ ,f 1 ],o 1 [,p 2 ,l 2 [ ,f 2 ],o 2 ]...)[,FORMAT= f]

 
 FILE = F1 
 F2 

 F1= ddname 
 F2= ddname 
 

Include all records from

 ALL  file in join processing
 NONE 
 
 INCLUDE    ,AND,  
 OMIT =  ,&,  
   (c   c ... ) 
 1  ,OR,  2 
  ,|,  
 

NOSEQCK Perform sequence check

when SORTED is speci-
fied

SORTED Presume records are

unsorted

TASKID=xx

VSAM files are processed

F  as fixed-length
TYPE=  
V 

Table 2. (Page 2 of 8)Control Statement Summary Chart

2–4 Syncsort MFX Programmer’s Guide

Control
Statement Parameters Delivered Default
Name

MERGE Century window starts

0  with current year
CENTWIN=  s 
f 

No checkpoint
CKPT
CHKPT

NOEQUALS
EQUALS
NOEQUALS

 FIELDS=(p 1 ,l 1 [ ,f 1 ],o 1 [,p 2 ,l 2 [ ,f 2 ],o 2 ]...)[,FORMAT=f] 

 
 FIELDS=COPY 

FILES=n

SKIPREC=n Copy all records

STOPAFT=n Copy all records

MODS No exits
 ,N 
 ,S 
 ,C 
exit-name 1 =(r 1 ,b 1 [,d 1 ]   ),...,exit-name 16 =(...)
 ,E 
 ,X 
 ,T 

OMIT Sort/Merge all records

 ALL 
 NONE 
 
  ,AND,  
COND=   ,&,  
 (c   c 2 ... ) [,FORMAT=f] 
 1 
 ,OR, 
  ,|,  
 

Table 2. (Page 3 of 8)Control Statement Summary Chart

Syncsort MFX Programmer’s Guide 2–5

Control
Statement Parameters Delivered Default
Name

OUTFIL ACCEPT=n

BLKCCH1

BLKCCH2

BLKCCT1

Record format unchanged

CONVERT
VTOF

ENDREC=n End processing with last

record

One output file

 fileid 
FILES =  
 ( fileid 1 [ ,fileid 2 ]... ) 

Output defined by FILES

 ddname 
FNAMES=  
 (ddname 1 [,ddname 2 ]...) 

FTOV Record format unchanged

No report heading
HEADER1=(field 1 [,field 2 ]...)

No page headings
HEADER2=(field 1 [,field 2 ]...)

Output all records

 INCLUDE   ALL 
  =  (comparisons) 
 OMIT   NONE 

IFTRAIL=(subparms)

60 (if report-writing
n  parameters)
LINES=  ANSI 
 (ANSI,n) 

NODETAIL Detailed report

Table 2. (Page 4 of 8)Control Statement Summary Chart

2–6 Syncsort MFX Programmer’s Guide

 Control
Statement Parameters Delivered Default
Name

OUTFIL Return code of zero

 RC0 
 
NOTMTOFL=  RC4 
 RC16 
 

Return code of zero

 RC0 
 
NULLOFL=  RC4 
 RC16 
 

OUTPUT = ( subparm )

Record Unchanged

   
   
 ,OUTREC 

 [ , PARSE=(subparm) ]  ,BUILD  
  = ( field [ , field ]... ) 
  1 2
  ,OVERLAY  
   
 
 , IFTHEN= ( subparm ) [ ,IFTHEN = ( subparm )... ] [ ,IFOUTLEN =n ]
 
 , FINDREP=(subparm) 

REMOVECC Produce a report with

ANSI control characters

REPEAT=n Records are not repeated

All records are produced

n  in output
SAMPLE=  
 (n,m) 

SAVE Omitted records not

saved for output

No sections
SECTIONS=(field 1 [,field 2 ]...)

SPLIT No split output

SPLIT1R=n No split output

SPLITBY=n No split output

Table 2. (Page 5 of 8)Control Statement Summary Chart

Syncsort MFX Programmer’s Guide 2–7

Control
Statement Parameters Delivered Default
Name

OUTFIL STARTREC=n Start processing with

first record

No report trailer
TRAILER1=(field 1 [,field 2 ]...)

No page trailers
TRAILER2=(field 1 [,field 2 ]...)

VLFILL=f Missing fields will be

filled with blanks (x'40')
when CONVERT option
in use; missing fields
cause application termi-
nation when CONVERT
is not specified

VLTRAIL=string No string appended to

records

Retain all trailing bytes

VLTRIM=b

OUTREC IFTHEN=(subparm)[,IFTHEN=... ] [ ,IFOUTLEN=n ]

 FIELDS   
[ PARSE=(subparm), ]  =(fields)  ,CONVERT 
 BUILD   ,VTOF 

[ PARSE=(subparm), ] OVERLAY=(fields)

FINDREP=(subparm)

RECORD
LENGTH=(l 1 ,...,l 7 )

F 
TYPE=  
V 

Table 2. (Page 6 of 8)Control Statement Summary Chart

2–8 Syncsort MFX Programmer’s Guide

Control
Statement Parameters Delivered Default
Name

REFORMAT
[ ,FILL=f ]

FIELDS= ( Fn:p 1 ,l 1 [ ,[Fn:]p 2 ,l 2 ] [ ,? ] [ ,[Fn:]p 3 ,l 3 ]... [ ,[Fn:]p m ] [ ,Fn:p n ] )

Table 2. (Page 7 of 8)Control Statement Summary Chart

Syncsort MFX Programmer’s Guide 2–9

Control
Statement Parameters Delivered Default
Name

SORT Century window starts at

0  current year
CENTWIN=  s 
f 

No checkpoint
CKPT
CHKPT

d 
 
  ( nn,mm )  
DYNALLOC =  (d,n (,RETRY=  OFF  [ ,SC=s ]) 
   
 
 OFF 

NOEQUALS
EQUALS
NOEQUALS

 FIELDS=(p 1 ,l 1 [ ,f 1 ],o 1 [,p 2 ,l 2 [ ,f 2 ],o 2 ]...)[,FORMAT=f] 

 
 FIELDS=COPY 

n 
FILSZ =  
 En 

n 
SIZE =  
 En 

SKIPREC=n Sort or copy all records

STOPAFT=n Sort or copy all records

SUM No summing of fields; no

 FIELDS=(p 1 ,l 1 [ ,f 1 ][,p 2 ,l 2 [ ,f 2 ] ] ...)[,FORMAT=f]  reduction of equal-keyed
 FIELDS=NONE  records
 

XSUM

Table 2. (Page 8 of 8)Control Statement Summary Chart

2–10 Syncsort MFX Programmer’s Guide

Disk Sort, MAXSORT, and PARASORT Control Statement
Requirements
The following table summarizes control statement usage for Disk Sort, MAXSORT,
and PARASORT.

Control Statement Disk Sort MAXSORT and PARASORT

ALTSEQ Optional Optional

DUPKEYS Optional; not applicable to copy Optional

END Required if exits included in input Required for MAXSORT if exits

stream included in input stream; optional for
PARASORT

INCLUDE/OMIT Optional Optional

INREC Optional Optional

JOIN Optional Not supported

JOINKEYS Optional Not supported

MERGE Required for merge or copy Not applicable

MODS Required for exits Required for exits

OUTFIL Required for multiple output or Optional

reports

OUTREC Optional Optional

RECORD Conditionally required Conditionally required

REFORMAT Optional Not supported

SORT Required for sort or copy Required for sort; copy not supported

SUM Optional; not applicable to copy Optional

Table 3. Control Statement Usage for Disk Sort, MAXSORT, and PARASORT

Syncsort MFX Programmer’s Guide 2–11

Data Utility Processing Sequence
The following figure illustrates the sequence in which MFX control statements and
parameters are processed. It includes those control statements and parameters
that modify the input file (e.g., INCLUDE/OMIT), reposition record fields (e.g.,
INREC, OUTREC), and create reports (e.g., OUTFIL).
When specifying record fields on any of these MFX control statements or
parameters, refer to the record as it appears at that stage of MFX processing. For
example, when specifying SORT fields be sure to take into account any
repositioning of fields that may be due to INREC processing.

2–12 Syncsort MFX Programmer’s Guide

 Input File Input File Input File
1 2

Record Selection for Record Selection for

Join File 1 Join File 2
INCLUDE/OMIT INCLUDE/OMIT
Parameters Parameters

Join Processing
JOIN, JOINKEYS, REFORMAT
Control Statements
Record Selection
Conventional SORTIN INCLUDE/OMIT
SORTJNF1, SORTJNF2
Sort Processing Control Statement
Join Processing

Field Selection
INREC
Control Statement

Record Arrangement
SORT
Control Statement

Combining/Eliminating Duplicate
Records
SUM or DUPKEYS
Control Statement

Printable and Easy to Read Output

and Variable to Fixed Length
Format Conversion
OUTREC
Control Statement

Record Selection for Record Selection for

Output File 1 Output File n
STARTREC, ENDREC, STARTREC, ENDREC,
INCLUDE/OMIT, SAVE INCLUDE/OMIT, SAVE
Parameters Parameters

Report Formatting Report Formatting

for for
Output File 1 Output File n

Printable & Easy to Read Printable & Easy to Read

Output for File 1 & Variable to Output for File 1 & Variable to
Fixed Length Format Conversion or Fixed Length Format Conversion or
Fixed to Variable Format Fixed to Variable Format
Conversion Conversion
OUTREC, CONVERT, FTOV OUTREC, CONVERT, FTOV
Parameters Parameters

Multiple Output and Report

Output File Formatting
1 OUTFIL Output
Control Statement(s) File n

Figure 1. Data Utility Processing Sequence

Syncsort MFX Programmer’s Guide 2–13

Control Statement Examples
Simple examples illustrating the syntax of each of the MFX control statements are
included in this chapter. More complex applications are presented in “Chapter 3,
How to Use the MFX Data Utility Features”. These applications demonstrate how
the INCLUDE/OMIT, INREC, JOIN, JOINKEYS, OUTFIL, OUTREC,
REFORMAT, and SUM control statements can be used to accomplish a variety of
tasks, such as selecting input records, selecting input fields, joining records with
matching keys, combining records, reformatting output records, writing reports,
and creating multiple output.

Rules for Control Statements

The following rules apply to MFX control statements.

Specifying Control Statements

• Control statements can be in any order, except for the END control statement
which, if specified, must be last.
• Each control statement, except for JOINKEYS and OUTFIL, can be specified
only once for a particular application.
• The control statement can begin in column 2 through column 69. If labels are
used, the control statement must be separated from the label by at least one
blank.
• The control statement name must be the first field (or the first field after a
label) of the first card image of the control statement. It cannot be continued on
a continuation card image.
• The last operand of each control statement must be followed by at least one
blank.

Specifying Parameters
• Parameters can take three forms:
• Parameter
• Parameter=value
Parameter=(value)
Parameter(value)
• Parameter=(value1,value2,...,valuen)
Parameter(value1,value2,...,valuen)
Note that multiple values must be enclosed in parentheses.

2–14 Syncsort MFX Programmer’s Guide

 • Parameters can be in any order, but if parameters are present, the first
parameter must begin on the first card image of a control statement.
• Parameters must be separated from each other by commas.
• The parameter(s) must be preceded and followed by at least one blank. A blank
separates the parameter(s) from the control statement name and also indicates
the end of the control statement.
• If the parameter(s) end in column 71, column 72 must contain a blank to signal
the end of the control statement.
• With the exception of literal strings and constants, a parameter value cannot
exceed 28 alphanumeric characters. Parameter values cannot include commas,
equal signs, or parentheses.
• With the exception of literal strings specified as parameter values, blanks are
not permitted within parameters.

Specifying Field Positions, Lengths, and Formats

• Control statements reference fields by position p and length l.
• The first byte of every fixed-length record is position 1, the second byte position
2, and so on.
• Bytes 1 through 4 of variable-length records are reserved for the Record
Descriptor Word (RDW). For these records, the first byte of the data portion is
position 5.
• Some control statements support bit-level processing. This means a binary
control field can begin and end on any bit of any byte. The 8 bits in each byte are
numbered 0 through 7. For example, a position value of 7.4 designates a field
beginning on the fifth bit of the seventh byte. A length value of 7.4 designates a
field 7 bytes, 4 bits long.
• Make sure the position value takes into account any record reformatting and
data conversion that may have resulted from MFX data utility processing or
exit programs. Refer to Figure 1 (“Data Utility Processing Sequence”) on page 2-
13 and “Chapter 8 The Flow of the Sort”.
• When proper processing depends on data format, the format of the field must be
specified.
• The format of the field must be appropriate to the task. For example, only
numeric fields can be SUMmed.
• When all the fields have the same format, the format value can be specified just
once through the FORMAT=f subparameter. The FORMAT=f subparameter
cannot be used when the INCLUDE/OMIT parameter is specified on the
OUTFIL control statement.

Syncsort MFX Programmer’s Guide 2–15

Specifying Comments
• Identify a comment card image by placing an asterisk (*) in column 1.
Comments can extend through column 80.
• To add a comment to a control statement card image, leave one or more blanks
after the last parameter or comma on the image and follow with the comment,
which can extend through column 71.
• Continue a comment that follows a control statement by coding an asterisk (*)
in column 1 of the next card image or, if the control statement had ended, by
placing a continuation character in column 72.
• Comment lines can be inserted between a control statement and its
continuation by coding an asterisk (*) in column one.

Specifying Continuation Card Images

Control statements cannot extend beyond column 71, but they can be continued. To
continue a control statement:
• Break after a parameter-comma or parameter-colon combination before column
72. Begin the continuation of the next card image anywhere between columns 2
and 71 if there is no label on the continuation card. If there is a label, begin the
continuation card in any column from 3-71. No continuation character is
required.
--or--

• When the control statement extends through column 71 and cannot be broken
at a parameter-comma or parameter-colon combination:
• If the control statement does not contain a literal string that would extend
beyond column 71, place a continuation character in column 72 and
continue the control statement on the next card image anywhere between
columns 2 and 71.
• If the control statement does contain a literal string that would extend
beyond column 71, place a continuation character in column 72 and begin
the continuation of the literal string in column 16 of the next card image.
The following examples illustrate how card images can be continued.

COL. 72
↓
SORT FIELDS=(1,10,A,20,5,A,45,7,A),FORMAT=CH,STOPAFT=100,
EQUALS

Figure 2. Continuing a Control Statement Without Specifying a Continuation Character

In the above example, no continuation character is required. The control statement

is interrupted after a parameter-comma combination before column 72.

2–16 Syncsort MFX Programmer’s Guide

 COL. 16 COL.72
↓ ↓
OUTFIL OUTREC=(1:10,8,30:40,10),HEADER2=(1:'CUSTOMER NUMBX
ER',30:'ITEM NUMBER')

Figure 3. Continuing a Control Statement with a Continuation Character

In this example, a continuation character is necessary because the literal string in

the HEADER2 specification would extend beyond column 71. The 'X' in column 72
is the continuation character. The literal string is continued in column 16 of the
next card image.

Specifying Labels
MFX supports labels. If labels are used, the following rules apply:
• Labels are permitted on all SYSIN control statements, including continuation
card images, but not on the control statements passed by an invoking program
or the $ORTPARM DD statement.
• Labels must begin in column 1 with an alphabetic character.
• Labels can be any length, provided the other rules which apply to control
statements are followed.
• At least one blank must separate the label from the control statement name or
parameter that follows it.

Notational Conventions Used in the MFX for z/OS Programmer’s

Guide
• Braces { } indicate that a choice must be made from the alternatives listed.
• Brackets [ ] indicate an optional item. Two or more vertically listed items in
brackets are mutually exclusive options; only one can be chosen for a particular
application.
• Defaults are underlined.
• Upper-case letters, numbers, commas, equal signs, and parentheses ( ) must be
entered exactly as indicated. Lower-case letters represent variables which must
be replaced by actual values.
• Subscripts show position in a series, and three dots indicate an ellipsis.
For example, a1,a2,...,a5 is equivalent to a1,a2,a3,a4,a5 and represents five a
items (variables which will be replaced with actual values).
• Examples that are to be entered exactly as shown are presented in the Courier
typeface, for instance:

Syncsort MFX Programmer’s Guide 2–17

 ALTSEQ CODE=(F0B7,F1B8,F2B9,F3BA,F4BB,F5BC,F6BD,F7BE,F8BF,F9C0)

Figure 4. Sample Examples in Courier Typeface

2–18 Syncsort MFX Programmer’s Guide

 ALTSEQ

ALTSEQ Control Statement

The ALTSEQ control statement constructs an alternate collating sequence for all
control fields for which the format code AQ has been specified. AQ can be specified
in the following locations:
• SORT/MERGE control statement
• INCLUDE/OMIT control statement
• JOINKEYS control statement
• INCLUDE/OMIT parameter on OUTFIL and JOINKEYS control
statements
• WHEN subparameter of IFTHEN on INREC, OUTREC and OUTFIL
control statements
If an alternate collating sequence has been provided by installation default, AQ
fields collate against this sequence, modified by the ALTSEQ control statement. If a
default alternate sequence has not been provided, AQ fields collate against the
standard EBCDIC sequence, modified by the ALTSEQ control statement. AQ can
be specified for one or more control fields so that those control fields all use the
same alternate collating sequence.
The ALTSEQ control statement also constructs an alternate collating sequence for
all control fields processed by the TRAN parameter of the INREC and OUTREC
control statements, as well as the TRAN subparameter of the OUTREC parameter
on the OUTFIL control statement.

ALTSEQ Control Statement Format

The format of the ALTSEQ control statement is illustrated below:

ALTSEQ CODE=(ccpp1[,ccpp2]...)

Figure 5. ALTSEQ Control Statement Format

CODE Parameter (Required)

The CODE parameter specifies how the characters of the current collating
sequence are to be reordered to create the alternate collating sequence.
The CODE parameter can contain from 1 to 256 entries, each consisting of four
hexadecimal digits. These entries must be separated by commas and enclosed in
parentheses.
Each CODE entry consists of two parts:
cc The cc value represents the character that is to be repositioned in the
alternate sequence.

Syncsort MFX Programmer’s Guide 2–19

 ALTSEQ

pp The pp value indicates where the character represented by the cc value

is to be repositioned in the alternate sequence.
The character represented by the cc value does not replace the character
represented by the pp value. If both characters occur as sort control fields, they will
be considered equal in the collating process.
Each character (cc entry) can be moved only one time. However, more than one cc
entry can be mapped to the same pp value.

Sample ALTSEQ Control Statements

ALTSEQ CODE=(F0B7,F1B8,F2B9,F3BA,F4BB,F5BC,F6BD,F7BE,F8BF,F9C0)

Figure 6. Sample ALTSEQ Control Statement

This sample ALTSEQ control statement shows that the numbers 0 through 9 are to
collate before the uppercase alphabet.

ALTSEQ CODE=(F040)

Figure 7. Sample ALTSEQ Control Statement

This sample ALTSEQ control statement specifies that the number 0 is to collate as
equal to a blank (X'40').

2–20 Syncsort MFX Programmer’s Guide

 DUPKEYS

DUPKEYS Control Statement

The DUPKEYS control statement is used to enable special processing for records
with equal sort/merge control fields (keys). You can perform the following functions:
• Sum specified numeric fields, place the sum in one record and delete the
other records with the same key (SUM)
• Compute the average of specified numeric fields, place the average in one
record and delete the other records with the same key (AVG)
• Determine the minimum or maximum value of specified numeric fields,
place this value in one record and delete the other records with the same
key (MIN,MAX)
• Delete all but one of the records with equal keys (FIELDS=NONE)
• Retain only records with keys that occur more than once (ALLDUPS)
• Retain only the first record of those with keys that occur more than once
(FIRSTDUP)
• Retain only the last record of those with keys that occur more than once
(LASTDUP)
• Retain only the records with keys that occur only once (NODUPS)
Note that the retained record will not necessarily be the first record unless the
EQUALS parameter is in effect.
The records deleted by DUPKEYS can optionally be written to a separate file.
The DUPKEYS control statement cannot be used with a SUM control statement,
nor when FIELDS=COPY is specified on the SORT or MERGE control statement.
If you need to add other DUPKEYS functionality to an application with a SUM
control statement, you must move the SUM specification to the DUPKEYS
statement and remove the SUM statement. If XSUM was used, then XDUP should
be specified and the JCL changed from using a SORTXSUM DD to a SORTXDUP
DD.

Syncsort MFX Programmer’s Guide 2–21

DUPKEYS

DUPKEYS Control Statement Format

The format of the DUPKEYS control statement is illustrated below.

 
 function [ ,function ]... [ ,FORMAT=f ] 
 FIELDS=NONE 
 ALLDUPS 
DUPKEYS   [ ,XDUP ]
 FIRSTDUP [,NODUPS] 
 LASTDUP [,NODUPS] 
 NODUPS 
 

where function is:

 AVG 
 MAX 
  = ( p 1 ,l 1 [ ,f 1 ] [ ,p 2 ,l 2 [ ,f 2 ] ]... )
 MIN 
 SUM 

Figure 8. DUPKEYS Control Statement Format

Function Parameters (AVG, MAX, MIN, SUM)

Each field specified in the AVG, MAX, MIN, and SUM parameters is identified by
its position p, length l, and format f, described as follows:
p The position value indicates the first byte of the field rela-
tive to the beginning of the input record after INREC
and/or E15 processing, if specified, have completed. The
field must begin on a byte boundary.
l The length value indicates the length of the field. The
length must be an integral number of bytes. Refer to Table
4 on page 2-23 for the permissible lengths.
f The optional format value indicates the data format. The
formats that can be specified are in Table 4 on page 2-23.
If all the defined fields have the same format, you can
specify the format value once by using the FORMAT=f
subparameter. If you specify both the individual f values
and the FORMAT subparameter, the individual f values
will be used for fields where they are specified.

2–22 Syncsort MFX Programmer’s Guide

 DUPKEYS

PERMISSIBLE LENGTH
FORMAT
CODE
SUM Fields MIN or MAX Fields AVG

BI 2, 4, or 8 bytes 1 to 256 bytes 2, 4, or 8 bytes

FD* 4, 8, or 16 bytes 4, 8, or 16 bytes 4, 8, or 16 bytes

FI 2, 4, or 8 bytes 2, 4, or 8 bytes 2, 4, or 8 bytes

FL 4, 8, or 16 bytes 4, 8, or 16 bytes 4, 8, or 16 bytes

PD 1 to 16 bytes 1 to 16 bytes 1 to 10 bytes

ZD 1 to 31 bytes 1 to 31 bytes 1 to 18 bytes

Note: *A non-finite number in the data will cause a WER497A error.

Table 4. Allowed DUPKEYS Formats and Field Lengths

AVG Parameter (Optional)

Use the AVG parameter to specify numeric fields to contain the average value
calculated from all records with the same control fields. Multiple fields separated
by commas may be specified in the same parameter. The results of the AVG
parameter will be truncated for all data formats except FL.
If overflow or underflow occurs during AVG calculations, the duplicate-keyed
records will not be deleted and none of the AVG, MAX, MIN, or SUM functions will
be performed.
Adding AVG to an existing MAXSORT application could cause the generation of
additional intermediate output files (SORTOU00 or SORTOUnn). This occurs
because AVG delays DUPKEYS processing until the final MAXSORT merge pass.

MAX Parameter (Optional)

Use the MAX parameter to specify numeric fields to retain the maximum value
among all records with the same control fields. Multiple fields separated by
commas may be specified in the same parameter. Equally-keyed records are
processed pair by pair. For the MAX parameter, the values in the MAX fields are
compared; the highest MAX value is placed in one of the records, and the other
equally-keyed records are deleted. The sorted data will be reduced to one record per
sort key value.

MIN Parameter (Optional)

Use the MIN parameter to specify numeric fields to retain the minimum value
among all records with the same control fields. Multiple fields separated by
commas may be specified in the same parameter. Equally-keyed records are

Syncsort MFX Programmer’s Guide 2–23

DUPKEYS

processed pair by pair. For the MIN parameter, the values in the MIN fields are
compared; the lowest MIN value is placed in one of the records, and the other
equally-keyed records are deleted. The sorted data will be reduced to one record per
sort key value.

SUM Parameter (Optional)

Use the SUM parameter to specify numeric fields to contain the summed value for
all records with the same control fields. Multiple fields separated by commas may
be specified in the same parameter. Equally-keyed records are processed pair by
pair. For the SUM parameter, the values in the SUM fields are added, the sum is
placed in one of the records, and the other record is deleted. The sorted data will be
reduced to one record per sort key value if arithmetic overflow does not occur
during the summing process.
If the sum of any of the specified SUM fields in any two equally-keyed records
overflows the size of the field, the duplicate-keyed record will not be deleted and
none of the AVG, MAX, MIN, or SUM functions will be performed.

FIELDS Parameter (Optional)

The only valid value for FIELDS is NONE. Specify FIELDS=NONE only if no
arithmetic functions are desired. The sorted data will be reduced to one record per
sort key value.

ALLDUPS Parameter (Optional)

Use the ALLDUPS parameter to specify that only records with sort/merge fields
that occur more than once are retained.

FIRSTDUP Parameter (Optional)

Use the FIRSTDUP parameter to specify that only the first record of those with
sort/merge fields that occur more than once should be retained. If the NODUPS
parameter is also specified, all records with sort/merge fields that occur exactly
once are also retained. For a merge, the first record will be the first record from the
lowest numbered input file.

LASTDUP Parameter (Optional)

Use the LASTDUP parameter to specify that only the last record of those with
sort/merge fields occurring more than once should be retained. If the NODUPS
parameter is also specified, all records with sort/merge fields occurring exactly once
are also retained. For a merge, the last record will be the last record from the
highest numbered input file.

2–24 Syncsort MFX Programmer’s Guide

 DUPKEYS

NODUPS Parameter (Optional)

Use the NODUPS parameter to specify that only records with sort/merge fields
that occur exactly once are retained.

XDUP Parameter (Optional)

Specify the XDUP parameter if you want records deleted by DUPKEYS processing
to be written to a data set defined by the SORTXDUP DD statement. These records
will be written to SORTXDUP at the time of DUPKEYS processing. The records
will not undergo OUTREC, E35, and OUTFIL processing because such processing
occurs after DUPKEYS processing.
The DCB BLKSIZE of the SORTIN data set will not be used to determine the
BLKSIZE of the SORTXDUP data set. System determined blocksize will be used
when enabled and appropriate. Unblocked output will be generated if system
determined blocksize has been disabled and an explicitly specified blocksize has not
been provided in the JCL.
The XDUP file will be sequenced in the same order as the SORTOUT file.
Note that XDUP may increase system requirements:
• Adding XDUP to an existing sort application may result in an increase in the
amount of SORTWORK space required. This occurs because XDUP delays all
DUPKEYS processing until Phase 3.
• Adding XDUP to an existing MAXSORT application could cause the generation
of additional intermediate output files (SORTOU00 or SORTOUnn). This
occurs because XDUP delays DUPKEYS processing until the final MAXSORT
merge pass.
• XDUP may require additional main memory. Specify a region size of 512K or
more.

General Considerations for DUPKEYS

• If NOEQUALS is in effect, the record which is retained during arithmetic
processing (AVG, MAX, MIN, SUM), or FIELDS=NONE processing, is
determined arbitrarily. If EQUALS is in effect, the record which is retained is
the first record read in a SORT application; in a MERGE, the retained record
will be from the lowest-numbered input file. The EQUALS parameter can be
specified on the SORT or MERGE control statement or as a PARM option.
• Functions (AVG, MAX, MIN, SUM), FIELDS=NONE, ALLDUPS, FIRSTDUP,
LASTDUP and NODUPS are all mutually exclusive parameters, except that
NODUPS can be specified with FIRSTDUP and LASTDUP.
• AVG, MAX, MIN, or SUM arithmetic cannot be performed on a SORT or
MERGE control field. An AVG, MAX, MIN, or SUM field cannot include any or
part of a SORT or MERGE control field.

Syncsort MFX Programmer’s Guide 2–25

DUPKEYS

• AVG, MAX, MIN, and SUM fields may not overlap each other.
• Each AVG, MAX, MIN and SUM parameter may be used only once.
• If any variable-length record does not contain all of the AVG, MAX, MIN, or
SUM fields, none of the arithmetic functions will be performed for that record.
• Non-AVG, non-MAX, non-MIN, and non-SUM fields remain unchanged and are
retained from the record which contains the average, maximum, minimum, or
sum value, respectively.
• If overflow or underflow occurs during AVG or SUM calculations for records,
then those records will not have any functions performed and none of the
records will be deleted. MAX and MIN calculations are also suspended among
those records. AVG, MAX, MIN, and SUM arithmetic restarts when a
subsequent set of records with equal control fields can be averaged or summed
without overflow. Further processing is determined by the option selected at
installation through the SUMOVFL parameter or the run-time parameter
OVFLO. If the RC16 option of this parameter has been selected, processing will
terminate with a WER049A critical error. For the RC0 (the delivered default) or
the RC4 option, average or sum processing will continue and a WER049I
message will be issued (only for the first occurrence). If a subsequent pair of
records with equal control fields can be averaged or summed without causing
overflow or underflow, the arithmetic functions will be performed. To avoid
arithmetic overflow with SUM, use the INREC control statement to insert zeros
of the proper format immediately before the SUM field. For example, for a PD
field, use nZ to insert binary zeros.
• Remember that the first 4 bytes of variable-length records are reserved for the
Record Descriptor Word, so the first byte of the data portion of the record is byte
5.
• DUPKEYS is incompatible with an incore sort. If you specify the DUPKEYS
control statement, allocate SORTWKxx data sets in the JCL or use the
DYNALLOC feature for dynamic SORTWK allocation. If no JCL SORTWKs are
provided and DYNALLOC is disabled by default, DUPKEYS will cause
DYNALLOC to be enabled.
• When AVG and SUM arithmetic is performed on FL fields, user-issued SPIE
macros are not permitted and exit routines must not produce exponent overflow
or underflow. Because of the numeric rounding performed by the hardware, the
exact average or sum depends on the order in which fields are calculated. Thus,
the average or sum may vary slightly for different executions.
• By default, the sign byte of a positive averaged or summed ZD field will be
converted to printable format. If you want to disable this action, use the
NZDPRINT PARM option. Refer to “ZDPRINT” on page 5-33.
• Adding ALLDUPS, FIRSTDUP, LASTDUP or NODUPS to an existing sort
application may result in an increase in the amount of SORTWORK space

2–26 Syncsort MFX Programmer’s Guide

 DUPKEYS

required. This occurs because these functions delay all DUPKEYS processing
until Phase 3.
• Adding ALLDUPS, FIRSTDUP, LASTDUP or NODUPS to an existing
MAXSORT application could cause the generation of additional intermediate
output files (SORTOU00 or SORTOUnn). This occurs because these functions
delay DUPKEYS processing until the final MAXSORT merge pass.

Sample DUPKEYS Control Statement

The following DUPKEYS statement deletes records with equal control fields but
places arithmetic sum, minimum, maximum, and average values of some fields in
the retained record.

DUPKEYS SUM=(20,8,32,4,FI),MIN=(40,6),MAX=(48,6),
AVG=(56,5,PD,64,7,PD),FORMAT=ZD

Figure 9. Sample DUPKEYS Statement

When the control fields are equal, this sample statement sums the ZD field
beginning in byte 20 and the FI field beginning in byte 32; selects the minimum
value of the ZD field beginning in byte 40; selects the maximum value of the ZD
field beginning in byte 48; averages the PD field beginning in byte 56 and the PD
field beginning in byte 64; and then deletes the equal-keyed record.

Syncsort MFX Programmer’s Guide 2–27

END

END Control Statement

If present, the END control statement must be the last control statement. The
END control statement is required only when the control statements are not
followed by /* or by a job control statement (i.e., when including exits in the input
stream).
The END control statement has no parameters, but can contain comments if the
comments are preceded by at least one blank.

2–28 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

INCLUDE/OMIT Control Statement

The INCLUDE/OMIT control statement selects records from an input file based on
comparisons testing the contents of one or more fields within the record. A field can
be compared to a constant or to another field within the record. Furthermore, a
binary field may enter into comparisons that involve testing the individual bits in
the field. Only one INCLUDE/OMIT control statement can be specified for an
application, either as an INCLUDE or as an OMIT control statement.

Locale-Based Comparison Processing

MFX supports alternative sets of collating rules based on a specified national
language. The alternative collating applies to INCLUDE/OMIT (and OUTFIL
INCLUDE/OMIT) comparison processing as well as to SORT/MERGE processing.
A locale defines single and multi-character collating rules for a cultural
environment.
Locale-based INCLUDE/OMIT processing applies only to character (CH) fields and
character or hexadecimal constants compared to character fields. When LOCALE is
active, a CH to BI (or BI to CH) comparison is not allowed. The illegal comparison
will cause MFX to terminate with an error message.
For more information on locale-based processing, see “LOCALE” on page 5-18.

INCLUDE/OMIT Control Statement Format

The format of the INCLUDE/OMIT control statement follows.

Syncsort MFX Programmer’s Guide 2–29

INCLUDE/OMIT

 ALL 
 
 NONE 
 INCLUDE    ,AND,  
  COND=    
 OMIT   (c  ,&,  c ... ) [,FORMAT=f] 
 1  ,OR,  2 
  ,|,  
 

c represents a comparison. Each comparison has this format:

 
  ,EQ,  
  ,NE,  
  ,GT,   p ,l [,f ]  
2 2 2 
  [,f 1 ]  ,GE,   
     constant  
  ,LT,  
   ,LE,  
 
   ,EQ,  
  [,f 1 ]  ,NE,  L(constant 1 [,constant 2 ]...)  
    
 
    
     
    ,BO,    
    ,ALL,    
    
     
    ,BM,    
    ,SOME,    
    
     
    ,BZ,    
    ,NONE,    
   bit mask  
   
    ,BNO,    
    ,NOTALL,     
  
  
   
 ,BI    ,BNM,   
p 1 ,l 1     ,NOTSOME,     
  
  
   
    ,BNZ,   
    ,NOTNONE,     
  
  
   
   
   bit pattern   
   UC   
   
   ,EQ,   LC   
    MC   
   ,NE,   UN   
   
   LN   
 

 MN  


 
 
  constant  
    
 ,SS  ,EQ,   L(constant 1 [,constant 2 ]...)  
  ,NE,   pattern constant  
   
 
 ,CSF 
  
  ,FD   
  ,FS  ,EQ,  NUM 
 ,PD  ,NE,  
  
  ,ZD  
 

 
 ,EQ, 
 ,NE, 
 ,GT, 
&MULTIINDD   constant
 ,GE, 
 ,LT, 
 ,LE, 
 

 
&MULTIINDD  ,EQ,  L(constant 1 [,constant 2 ]...)
 ,NE, 

Figure 10. INCLUDE/OMIT Control Statement Format

2–30 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

COND Parameter (Required)

The COND parameter controls how records are included or omitted from an
application. There are three forms of the COND parameter:
COND=ALL All of the input records are to be included. This is the
default.
COND=NONE None of the input records are to be included.
COND=comparison(s) Specifies one or more comparisons that determine which
records are to be included or omitted. Two types of com-
parisons are possible:
• A standard comparison, between two record fields
or between a record field and a constant. A binary
input field also allows comparison by bit mask or bit
pattern.
• A substring comparison, which allows the search
for a constant within a field, or for a field value within
a constant, or for a pattern constant (wildcard) within
a field. Use SS as the format to indicate a substring
comparison.
The following several pages describe standard comparisons. For information on
substring comparisons, see “Substring Comparisons” on page 2-46.
Each field specified in the COND parameter is identified by its position (p), length
(l) and format (f). When processing variable-length records, by default all fields
specified must be contained within the record. If an application is expected to
reference fields not completely contained within the record, refer to “VLTEST” on
page 5-30. VLTESTI provides for processing of records that do not contain all fields.
p The position value indicates the first byte of the field relative to the
beginning of the input record after E15 or E32 processing, if specified,
has completed. The field must begin on a byte boundary. (Keep in mind
that if a variable-length file is being referenced, the first 4 bytes must be
reserved for the Record Descriptor Word.)
l The length value indicates the length of the field. The length must be an
integer number of bytes. See the table below for permissible field lengths
by format.
f The format value indicates the format of the field. The permissible for-
mats for standard comparisons are indicated in the following table. If all
data fields have the same format, the FORMAT=f subparameter can be
specified instead of the individual f values. If both are specified, the indi-
vidual f values will be used for fields where they are specified.

Syncsort MFX Programmer’s Guide 2–31

INCLUDE/OMIT

Acceptable Field
Code Data Format
Length (Bytes)

AC EBCDIC characters are translated to their ASCII equivalents before 1 to 256

merging.

AQ Character. Records are merged according to an alternate sequence spec- 1 to 256

ified either in the ALTSEQ control statement or as an installation
default.

ASL Leading separate sign. An ASCII + or - precedes numeric field. One digit 2 to 256
per byte.

AST Trailing separate sign. An ASCII + or - trails numeric field. One digit 2 to 256
per byte.

BI Binary. Unsigned. 1 to 256

CH Character. Unsigned. 1 to 256

CLO Leading overpunch sign. Hexadecimal F,C,E, or A in the first 4 bits of 1 to 256
OL your field indicates a positive number. Hexadecimal D or B in the first 4
bits indicates a negative number. One digit per byte. CMP=CLC is
forced.

CSF Floating sign format. An optional leading sign may be specified immedi- 1 to 32*
FS ately to the left of the digits. If the sign is a -, the number is treated as
negative. For other characters, the number is treated as positive. Char-
acters to the left of the sign are ignored.

CSL Leading separate sign. An EBCDIC + or - precedes numeric field. One 2 to 256
LS digit per byte. CMP=CLC is forced.

CST Trailing separate sign. An EBCDIC + or - follows numeric field. One 2 to 256
TS digit per byte. CMP=CLC is forced.

CTO Trailing overpunch in the first 4 bits of the rightmost byte gives the 1 to 256
OT sign. Hexadecimal F,C,E, or A indicates a positive number. Hexadecimal
D or B indicates a negative number. One digit per byte. CMP=CLC is
forced.

FI Fixed point. Signed. (Equivalent to Signed Binary.) 1 to 256

PD Packed decimal. Signed. 1 to 255

PD0 Packed decimal. 2-8-byte packed decimal data with the first digit and 2 to 8
trailing sign ignored. The remaining bytes are treated as packed deci-
mal digits. Typically PD0 is used with century window processing and
Y2P format; Y2P processes the year, while PD0 processes month and
day.

Table 5. (Page 1 of 3)Valid Formats and Lengths of Include/Omit Fields

2–32 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

Acceptable Field
Code Data Format
Length (Bytes)

SFF Signed free format. Decimal digits (0-9) are extracted from right to left 1 to 44
to form a number value. A character of – or ) found within the field will
cause the value to be treated as a negative number. All other non-deci-
mal digit values in the field are ignored. A maximum of 31 digits can be
provided. When more than 31 digits are found in the field, the leftmost
digits will be ignored.

SS Substring format. 1 to 32752

UFF Unsigned free format. Decimal digits (0-9) are extracted from right to 1 to 44
left to form a positive number value. All non-decimal digit values in the
field are ignored. A maximum of 31 digits can be provided. When more
than 31 digits are found in the field, the leftmost digits will be ignored.

Y2B Binary. 2-digit, 1-byte binary year data treated as a 4-digit year by 1
CENTWIN (century window) processing.

Y2C Character. 2-digit character year data treated as a 4-digit year by 2

CENTWIN (century window) processing. Processing is identical to Y2Z
fields.

Y2D Packed decimal. 2-digit, 1-byte packed decimal year data treated as a 4- 1
digit year by CENTWIN (century window) processing.

Y2P Packed decimal. 2-digit, 2-byte packed decimal year data. Of the four 2
packed digits contained in the 2 bytes, the first digit and trailing sign
are ignored; the two inner digits are treated as a 4-digit year by
CENTWIN processing.

Y2S Character or zoned decimal. 2-digit, 2-byte valid numeric data treated 2
as a 4-digit year by CENTWIN (century window) processing, as for Y2C
and Y2Z. However, certain data are not treated as year data. Data with
binary zeros (X'00') or a blank (X'40') in the first byte will be collated
before valid numeric year data for ascending order (after year data for
descending order). Data with all binary ones (X'FF') in the first byte will
be collated after valid numeric year data for ascending order (before
year data for descending order). Zones are ignored, as for Y2C and Y2Z,
except for data where the first byte begins with X'00', X'40' or X'FF'.

Y2T Full-date, character, binary, or packed decimal formats. Full-date data 2 to 6

formats can be used to merge a variety of date fields. They can process
Y2U dates ending or starting with year digits (x...xyy or yyx...x). They can
also process non-date data commonly used with dates.
Y2V

Y2W

Y2X

Y2Y

Table 5. (Page 2 of 3)Valid Formats and Lengths of Include/Omit Fields

Syncsort MFX Programmer’s Guide 2–33

INCLUDE/OMIT

Acceptable Field
Code Data Format
Length (Bytes)

Y2Z Zoned decimal. 2-digit, 2-byte zoned decimal year data treated as a 2
4-digit year by CENTWIN (century window) processing. The zones are
ignored. Processing is identical to Y2C fields.

ZD Zoned decimal. Trailing overpunch in the first 4 bits of the rightmost 1 to 256
byte gives the sign. Hexadecimal F,C,E, or A indicates a positive num-
ber. Hexadecimal D or B indicates a negative number. One digit per
byte. CTO forces CMP=CLC.

Note: *1 to 256 when used with the NUM subparameter.

Table 5. (Page 3 of 3)Valid Formats and Lengths of Include/Omit Fields

For information on the year data formats (Y2B, Y2C, Y2D, Y2P, Y2S and Y2Z) plus
the related data format PD0, see “CENTWIN Parameter (Optional)” on page 2-71
and “Converting Year Data with Century Window Processing on INREC, OUTREC,
or OUTFIL OUTREC” on page 2-162. Also, see “Specifying Field-to-Field Standard
Comparisons for Year Fields” on page 2-39.
The constant to which a field can be compared may be one of the following types:
decimal A decimal constant can be any length. It should not be enclosed in
single quotes. It may or may not include a leading + or - sign. For
example, 100 is a valid decimal constant. The following numeric
data compare as equal: +0, -0, 0. The &DATExP date parameter
represents the current date as a decimal number (+n) to which a
field can be compared. See page 2-43 for more details.
hexadecimal A hexadecimal constant should be preceded by an X and specified
in pairs of valid hexadecimal values which must be enclosed in
single quotes: X'hh...hh'. For example, X'ACBF05' is a valid hexa-
decimal constant. The sign of the field is implicit in the represen-
tation.
character A character constant should be preceded by a C and enclosed in
single quotes: C'literal'. For example, C'SALES' is a valid charac-
ter constant.
The &DATEx and &DATEx(c) date parameters represent the
current date as a character string (C'string') to which a field can
be compared. See page 2-43 for more details.
You can also include or omit records based on whether their dates
fall within a specified time frame before or after the current date.
See page 2-44 for more details.
To include an apostrophe in a character constant, specify it as
two apostrophes; for example, C'D''AGOSTINO'. If a character
constant must be continued on a second card image, place a con-

2–34 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

tinuation character in column 72 and then begin the continuation

of the constant in column 16 of the next card image.
There are two methods in which the bit level characteristics of a binary input field
can be used to include or omit records. One is to compare the binary field to a bit
mask; the other is to compare the binary field to a bit pattern.
bit mask A bit mask is a string of bits, specified in terms of either hexadec-
imal or binary digits. The bit mask indicates which bits in the
input field are to be tested. Each bit in the mask whose value is 1
(ON) is tested against the corresponding bit in the input field. If
the value of a mask bit is 0 (OFF), the corresponding bit in the
input field is ignored.
The hexadecimal format of a bit mask is X'hh...hh,' where each
'hh' represents any pair of hexadecimal digits.
The binary format of a bit mask is B'bbbbbbbb...bbbbbbbb', where
each 'bbbbbbbb' represents 8 bits or a byte. Each bit is 1 or 0. The
number of bits in a binary bit mask must be a multiple of 8. The
maximum length of a binary bit mask is 256 bytes (2048 bits).
A bit mask is truncated or padded on the right to the byte length
of the binary field. The pad character is X'00' or B'00000000'.
bit pattern The binary format of a bit pattern is B'bbbbbbbb...bbbbbbbb',
where each 'bbbbbbbb' represents 8 bits or a byte. Each bit is 1, 0,
or period (.). If the value of a bit in the bit pattern is 1 or 0, the
corresponding bit in the binary input field is compared to 1 or 0.
If a . (period) occurs in a bit position in the bit pattern, the corre-
sponding bit in the input field is ignored.
The number of bit positions in a bit pattern must be a multiple of
8. The maximum length of a bit pattern is 256 bytes (2048 bits).
A bit pattern is truncated or padded rightward to the byte length
of the binary input field. The pad character is B'00000000'.
The comparison operators represent the following conditions:
EQ Equal to
NE Not equal to
GT Greater than
GE Greater than or equal to
LT Less than
LE Less than or equal to
BO (or ALL) All mask bits are 1s (ON) in the input field
BM (SOME) Some but not all mask bits are 1s (ON) in the input field

Syncsort MFX Programmer’s Guide 2–35

INCLUDE/OMIT

BZ (NONE) None of the mask bits is 1 (ON) in the input field

BNO (NOTALL) Some or no mask bits are 1s (ON) in the input field
BNM (NOTSOME) All or no mask bits are 1s (ON) in the input field
BNZ (NOTNONE) All or some mask bits are 1s (ON) in the input field

Rules for Multiple Comparisons

Multiple comparisons are separated by ANDs or ORs to form a logical expression.
(Alternatively, & and | may be used for AND and OR). When evaluating an
expression, each comparison cn is evaluated first. Then, AND conditions are
evaluated before OR conditions.
Parentheses may be used around groups of comparisons to change the default
evaluation order. Any number of nested parentheses may be used. Conditions
within parentheses are evaluated first, from innermost to outermost parentheses.
For example, if you wanted to select all records from your Paris office for 1995 and
1996, you might incorrectly specify:
INCLUDE COND=(1,4,CH,EQ,C'1995',OR,1,4,CH,EQ,C'1996',
AND,5,5,CH,EQ,C'PARIS')
The AND operator in the above statement would be evaluated first, producing
unexpected output. The correct statement would be:
INCLUDE COND=((1,4,CH,EQ,C'1995',OR,1,4,CH,EQ,C'1996'),
AND,5,5,CH,EQ,C'PARIS')
The added parentheses force the OR operator to be evaluated first, thus producing
the expected output.

Simplified Expression of EQ/OR and NE/AND Conditions

INCLUDE/OMIT comparisons implementing EQ/OR and NE/AND conditions can be sim-
plified from having to restate the same field data (p,l,f) when comparing the field with more
than one constant.

Since INCLUDE/OMIT comparisons are often coded to compare one field in the record to a
long list of constants, this requires repeating the position p, length l and format f (option-
ally) of the field for each constant. For example,

INCLUDE COND=(1,2,CH,EQ,C'NY',OR,1,2,CH,EQ,C'NJ',
OR,1,2,CH,EQ,C'CT',OR,1,2,CH,...)
However, this statement can be simplified to
INCLUDE COND=(1,2,CH,EQ,L(C'NY',C'NJ',C'CT',...))
In the simplified statement above, the field data (1,2,CH) and comparison operator
EQ are stated only once; the compared constants are grouped together in

2–36 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

parentheses preceded by ‘L’ for ‘list’; and OR is implied by use of EQ in the

statement.
This simplified statement is only permitted when the comparison operator is EQ or
NE. If EQ is specified, the comparison conjunction OR is implied in the statement.
If NE is specified, the comparison conjunction AND is implied in the statement. All
constants that are compatible with the p,l,f data field are permitted in the
simplified constant list.

Using Data in a File as Comparison Constants

If you want to compare a field to a list of constants which reside in a file, MFX’s join
facility could be used instead of INCLUDE or OMIT.
For instance, if you needed to compare a field in a master file to a long list of airport
codes, you would normally need to create an INCLUDE or OMIT condition such as
1,3,CH,EQ,C’JFK’,OR,1,3,CH,EQ,C’LAX’,OR,1,3,CH,EQ,C’DFW’,OR,1,3,CH,EQ,C’
ATL’,...
(or the shorthand equivalent 1,3,CH,EQ,L(C’JFK’,C’LAX’,C’DFW’,C’ATL’,...)).
But if all the codes were in a second file in the form
....JFK..... (record 1)
....LAX..... (record 2)
....DFW.... (record 3)
....ATL..... (record 4)
................ ...
you could easily use the join facility to read the file directly and compare
a field in each record to a field in the master file. This technique would
perform the equivalent processing, eliminating the need for a lengthy
control statement.
See “Using Join Processing To Copy a Large Number of Master File Records” on
page 3-23 for an example of how to do this with MFX’s join facility.

Syncsort MFX Programmer’s Guide 2–37

INCLUDE/OMIT

Specifying Field-to-Field Standard Comparisons for Non-date Fields

The format of a data field determines whether or not it can be compared to another
data field. This table illustrates which field-to-field comparisons are permitted.

CL CS CS CS CT
AS AS PD SF UF
AC AQ BI CH O F L T O FI PD ZD
L T 0 F F
OL FS LS TS OT

AC X

AQ X

X X
AS
L

X X
AS
T

BI X X

CH X X

CL X X
O
OL

X X X X X
CS
F
FS

X X X X X
CS
L
LS

X X X X X
CS
T
TS

CT X X
O
OT

FI X

PD X X

Table 6. Permissible Field-to-Field Comparisons for Non-year Data Formats

2–38 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

CL CS CS CS CT
AS AS PD SF UF
AC AQ BI CH O F L T O FI PD ZD
L T 0 F F
OL FS LS TS OT

X
PD
0

SF X X X X X
F

UF X X X X X
F

ZD X X

Table 6. Permissible Field-to-Field Comparisons for Non-year Data Formats

Padding of Compared Fields

When two fields are compared, the shorter field is padded to the length of the
longer field. Padding takes place as follows:
• The padding characters are blanks when the shorter field is in character
format; otherwise, they are zeros of the shorter field’s own format.
• Padding is on the right if the shorter field is in BI, CH or PD0 formats. Padding
is on the left for all other formats.

Specifying Field-to-Field Standard Comparisons for Year Fields

The year data formats that can be used with INCLUDE/OMIT are Y2B, Y2C, Y2D,
Y2P, Y2S and Y2Z. Year data formats can only be compared to other year formats;
they cannot be compared to formats in the table above.
The full date formats that can be used with INCLUDE/OMIT are Y2T, Y2U, Y2V,
Y2W, Y2X, and Y2Y. The full date formats may only be compared to other 2-digit
year full date formats with the same number of non-year digits.
The year data formats work with the CENTWIN run-time parameter or
installation option to define a 2-digit year value that is to be treated as a 4-digit
year. CENTWIN defines a sliding or fixed 100-year window that determines the
century to which 2-digit year data belong when processed by INCLUDE/OMIT and
other control statements.
The year data formats and CENTWIN ensure that century evaluation is applied to
INCLUDE/OMIT comparison conditions involving 2-digit year data. For example,
without CENTWIN processing, an INCLUDE/OMIT comparison would treat the
year 01 as "less than" the year 98. With CENTWIN processing, the 01 field could be
recognized as a twenty-first century date (2001), which would be treated as
"greater than" 98 (1998).

Syncsort MFX Programmer’s Guide 2–39

INCLUDE/OMIT

For details on the CENTWIN option, see “CENTWIN” on page 5-6. For details on
the year data formats, see “CENTWIN Parameter (Optional)” on page 2-239. For an
example of an INCLUDE control statement with a condition involving a year data
field, see Figure 23 on page 2-53.
For any of the 2-digit year formats, it is valid to compare them with any of the other
formats. Specifically, Y2B, Y2C, Y2D, Y2P, Y2S, and Y2Z fields can be compared to
each other.
The following table summarizes the valid field to field comparisons for Full-Date
formats:

Date Form Length and Data Format Allowed

yyx and xyy 3,Y2T

3,Y2W

2,Y2U

2,Y2X

yyxx and xxyy 4,Y2T

4,Y2W

3,Y2V

3,Y2Y

yyxxx and xxxyy 5,Y2T

5,Y2W

3,Y2U

3,Y2X

yyxxxx and xxxxyy 6,Y2T

6,Y2W

4,Y2V

4,Y2Y

Table 7. Permissible Field-to-Field Comparisons for Full-Date Formats

2–40 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

Specifying Field-to-Constant Standard Comparisons

The format of a data field determines the type of constant to which it can be
compared. The figure below illustrates which field-to-constant comparisons are
permitted.

Binary Year
Format Decimal Hexadecimal Character
(bit pattern) Constant

AC X X

AQ X X

ASL X

AST X

BI X* X X X

CH X X

CLO/OL X

CSF/FS X

CSL/LS X

CST/TS X

CTO/OT X

FI X**

PD X

PD0 X

SS X X

Y2B X X

Y2C/Y2Z X X

Y2D X X

Y2P X X

Y2S X X

Y2T*** X X

Y2U*** X X

Y2V*** X X

Table 8. (Page 1 of 2)Permissible Field-to-Constant Comparisons

Syncsort MFX Programmer’s Guide 2–41

INCLUDE/OMIT

Binary Year
Format Decimal Hexadecimal Character
(bit pattern) Constant

Y2W*** X X

Y2X*** X X

Y2Y*** X X

SFF X

UFF X

ZD X

Note: * The decimal constant cannot be higher than 18446744073709551615 or lower than 0.
** The decimal constant cannot be higher than 9223372036854775807 or lower than
-9223372036854775808.
*** Full-Date formats

Table 8. (Page 2 of 2)Permissible Field-to-Constant Comparisons

2–42 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

A constant will be padded or truncated to the length of the field with which it is
compared. Decimal constants are padded or truncated on the left; hexadecimal,
binary, and character constants are padded on the right. The padding characters
are:
Binary string B'00000000'
Character string X'40'
Hexadecimal string X'00'
Decimal fields Zeros of proper format. Decimal constants for 2-digit
year formats are padded or truncated to two decimal
digits representing a year. The year constant will then
have CENTWIN processing applied to it for compari-
son to a Y2 field. These are only for the two digit year
fields, not for full date constants.
The constants for PD0 comparison should not include the first digit and trailing
sign of the PD0 data that will be ignored. Thus, a PD0 field of n bytes will be
compared to a constant of n-1 bytes.

Current Date Constant Specification

You can compare fields to the date of an MFX run or the date of the run with an
offset in addition to decimal fields and binary, character, and hexadecimal strings.
Thus, records can more easily be included or omitted based on whether their dates
are equal to, less than, or greater than the run date or the run date with an offset.
The format of a current date constant is illustrated below.
+ 
current date constant   nnnn
– 
Figure 11. Current Date Constant Format

where:
current date constant Specifies a form of one of the &DATEx, &DATEx(c),
&DATExP, or Y'DATEx' parameters where x is 1, 2, 3, 4 or
5 and depends on date comparison compatibility.
+ Specifies a date after the current date.
– Specifies a date before the current date.
nnnn Specifies the number of offset days or offset months
depending upon x in the following cases: when the x in
&DATEx, &DATEx(c), &DATExP, or Y'DATEx' is 1, 3, 4 or
5, ‘nnnn’ represents offset days and can be from 0-9999;
when the x in &DATEx, &DATEx(c), &DATExP, or

Syncsort MFX Programmer’s Guide 2–43

INCLUDE/OMIT

Y'DATEx' is 2, ‘nnnn’ represents offset months and can be

from 0-999.
For an example of an INCLUDE control statement that uses a date range based on
a date constant, see Figure 24 on page 2-53.
The forms of current date constants available for standard comparisons are:
• &DATEx and &DATEx(c) represent the current date as a character string
(C'string') to which a field can be compared.
• &DATExP represents the current date as a decimal number (+n) to which a
field can be compared.
• Y'DATEx' represents the current date with a Y constant (Y'string') to which
a field can be compared.
The following table shows the current date constants and the format produced by
each. The c character in &DATEx(c) represents a non-blank separator character,
except open and close parentheses.

Current Date Constant Generated Constant

&DATE1 [{±}nnnn] C'yyyymmdd'

&DATE1(c) [{±}nnnn] C'yyyycmmcdd'

&DATE1P [{±}nnnn] +yyyymmdd

&DATE2 [{±}nnn] C'yyyymm'

&DATE2(c) [{±}nnn] C'yyyycmm'

&DATE2P [{±}nnn] +yyyymm

&DATE3 [{±}nnnn] C'yyyyddd'

&DATE3(c) [{±}nnnn] C'yyyycddd'

&DATE3P [{±}nnnn] +yyyyddd

&DATE4 [{±}nnnn] C'yyyy-mm-dd-hh.mm.ss'

&DATE5[{±}nnnn] C'yyyy-mm-dd-hh.mm.ss.nnnnnn'

Y'DATE1' Y'yymmdd'

Y'DATE2' Y'yymm'

Y'DATE3' Y'yyddd'

Table 9. Current Date Constant Formats

2–44 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

Full-Date Format Constant Specifications

Constants used for full-date comparisons should have the same number of digits in
the constant as in the full-date field that has been specified. Leading zeros must be
specified when needed. The constant is constructed from two items; the first is a 2-
digit year and the second is a value representing the months or days that comprise
the remainder of the full date format. For example, if a 5-byte Y2W field were to be
compared for a value greater than the 20th day of 1996, 96020 should be the code
for the constant.
Constants can be coded to represent special values, such as those found in header
or trailer records. All zeros or nines may be used with Y2T, Y2U, Y2V, Y2W, Y2X,
and Y2Y. The same number of digits must be present as in the field that is being
compared. The constant string Y'LOW' (representing binary zeros), Y'HIGH'
(representing binary ones), or Y'BLANKS' (representing blanks) may be coded with
the fields Y2T, Y2W, and Y2S. Y'DATEx' (representing the current date) may be
coded with certain full-date formats specifically (see Table 10 on page 2-45).

Y Constant Date Form Length and Data Format Allowed

Y'DATE1' yyxxxx and xxxxyy 6,Y2T

6,Y2W

4,Y2V

4,Y2Y

Y'DATE2' yyxx and xxyy 4,Y2T

4,Y2W

3,Y2V

3,Y2Y

Y'DATE3' yyxxx and xxxyy 5,Y2T

5,Y2W

3,Y2U

3,Y2X

Table 10. Full-Date Comparisons

Syncsort MFX Programmer’s Guide 2–45

INCLUDE/OMIT

Substring Comparisons
Substring comparison (SS format) can be based on any of the following searches:
• Match occurrence of a constant within a record field
• Match occurrence of a record field within a constant
• Match occurrence of a pattern (wildcard) constant within a record field
In the first form, the length of the constant is less than the length of a specified
field. Records will be searched for the occurrence of the constant anywhere within
the field. The condition will be true if an EQ operator is specified and the constant
is found or if a NE operator was specified and the constant is not found. For
example, consider the constant “ANYTOWN” and a 60-byte field that contains an
address. Records will be searched for the occurrence of the literal “ANYTOWN”
anywhere within the 60-byte address field. If a match is found and the logical
operator is EQ, then the logical result is “true.” The logical result is also “true” if
the literal does not appear within the 60 bytes and the logical operator is NE.
In the second form, the length of a constant is greater than or equal to the length
of a specified field. Records will be searched for an occurrence of the field within the
constant. For example, the constant 'A02,A05,A06,A09' can be compared against
the contents of a 3-byte field within the record. This constant is composed of four 3-
byte substrings in a format that is known to match the data in the record field,
separated by commas, a character known not to be in the record field. This means
that ‘A02’, ‘A05’, ‘A06’, and ‘A09’ are the only possible 3-byte strings that could
possibly match the data in the record field. If the 3-byte field matches any 3-byte
character string in the constant, the logical result is “true” if the logical operator is
EQ. If the 3-byte field does not match all 3-byte character strings in the constant,
the logical result is “true” if the logical operator is NE.
The character used to separate elements of the constant should be a character that
does not appear in the field being compared. The comparison is then equivalent to a
standard comparison with ORed conditions when the logical operator is EQ. That
is, the condition is true if 'A02' OR 'A05' OR 'A06' OR 'A09' is found in the field
being compared. The substring comparison is a much more compact expression
than multiple OR conditions in a standard comparison. When the logical operator
is NE, the comparison is equivalent to a standard comparison with ANDed
conditions.
In the third form, a pattern constant is used to describe a string that is the object of
the search rather than the exact characters to be found. The pattern constant can
consist of one or more character or hexadecimal constants and wildcard characters.
A wildcard can be either a percent sign ‘%’ which matches any single character or
an asterisk ‘*’ which can match zero or more characters in a string.
In this form of searching, only an occurrence of a pattern constant within a record
field can be performed. An occurrence of a record field cannot be searched for in a
pattern constant.

2–46 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

Typically this form of substring search is used when the beginning and ending
characters of a field contain the major values of the search. The middle characters
represent a range of values within this major category that are desired for
selection.
For all forms of substring comparison, fields in the record can be from 1 to 32752
bytes in length. Constants can be in either character or hexadecimal format and
can be from 1 to 256 bytes in length. (See the description of constants just after
Table 8 on page 2-41.)
See examples starting with Figure 17 on page 2-51 for some samples of how to use
the substring comparison.

NUM Subparameter
Use NUM to identify a field as numeric or non-numeric in CSF/FS, FD, PD or ZD
format. Specify NUM with the field (p,l), format (CSF/FS, FD, PD or ZD) and
comparison operators (EQ or NE), described as follows:
p,l Specify field subparameters p as the starting position of the
field in the record; and l as the length in bytes (1 to 256) of the
field.

CSF/FS/FD/PD/ZD Specify CSF or FS format to evaluate the field for character

numerics. A field is character numeric if every byte contains
only characters from 0 to 9. In hexadecimal format, for
X'hhhh...hh', every 'hh' represents hex values in the range F0
to F9. Otherwise, the field is non-numeric. For example, 2468
and X'F3F2F1' are character numeric; 24A68 and X'F3F2C1'
are character non-numeric.
Specify PD format to evaluate the field for packed decimal
numerics. A field is packed decimal numeric if every non-sign
byte contains values from the range of 0 to 9 and the sign byte
is C, D, or F. In hexadecimal format, for X'ppp...ps', every p
represents values in the range 0 to 9; s represents C, D, or F.
Otherwise, the field is non-numeric. For example, X'2468C'
and X'1359D' are packed decimal numeric; X'24A68' and
X'F3F2B1' are packed decimal non-numeric.
Specify ZD format to evaluate the field for zoned decimal
numerics. A field is zoned decimal numeric if every non-sign
byte contains only characters from 0 to 9 and the sign byte
contains hex values in the ranges of C0 to C9, D0 to D9, or F0
to F9. In hexadecimal format, for X'zzzz...sz', every zz rep-
resents hex values in the range of F0-F9; sz represents hex
values in the ranges of C0 to C9, D0 to D9, or F0 to F9. Other-
wise, the field is non-numeric. For example, 2468 and

Syncsort MFX Programmer’s Guide 2–47

INCLUDE/OMIT

X'F3F2F1D5' are zoned decimal numeric; 24A68 and

X'F3F2B1' are zoned decimal non-numeric.
Specify FD format to evaluate the field for decimal floating
point numerics. A field is numeric if it is a finite number. Pos-
itive, negative infinity and signaling and quiet NaNs (Not-a-
Number) are considered non-numeric.
EQ/NE Specify EQ to evaluate the field for numerics; specify NE to
evaluate the field for non-numerics.
See Figure 21 on page 2-52 for an example of how to use the NUM subparameter.

Alphanumeric Comparisons
A field in a record may be tested to see if it is upper-case alphabetic (UC), lower-
case alphabetic (LC), mixed case alphabetic (MC), upper-case alphabetic and
numeric (UN), lower-case alphabetic and numeric (LN) or mixed-case alphabetic
and numeric (MN).
Upper-case alphabetic is defined as A–Z, lower-case alphabetic is defined as a
through z, and numeric is defined as 0 through 9.
The field in the record must be BI format. The comparison operator can be EQ or
NE.
For example,
INCLUDE COND=(10,4,BI,EQ,LN)
will include records where the field in column 10 for 4 bytes contains only lower-
case alphabetic characters and/or numeric characters.

&MULTIINDD
&MULTIINDD is used together with the MULTIIN PARM which directs MFX to
read multiple input files from separate DD statements. &MULTIINDD is defined
as the two-byte C’nn’ of the SORTMInn ddname (or the two-byte C’n ’ for
SORTMIn) from which a record was read.
&MULTIINDD is used in place of p,l,CH to compare to a two-byte character string
to determine the input record’s origin.
If the MULTIIN PARM option is not in effect, then &MULTIINDD is defined as two
blanks (C' '). If an E15 or E32 exit inserts a record, &MULTIINDD will be C’EX’ for
that record. There will be no change to the &MULTIINDD definition when an E15
exit accepts or changes a record.
&MULTIINDD can only be specified on the INCLUDE/OMIT statement.
See Figure 22 on page 2-52 for an example of how to use &MULTIINDD.

2–48 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

Sample INCLUDE/OMIT Control Statements

Example 1

INCLUDE COND=(24,4,PD,LT,28,4,PD,OR,10,2,CH,EQ,C'NY')

Figure 12. Sample INCLUDE Control Statement

In this example, records will be included in the application if the numeric value in
the field beginning in byte 24 is less than the numeric value in the field beginning
in byte 28 or if the character value in the field beginning in byte 10 is equal to NY.
Example 2

OMIT COND=(1,3,ZD,EQ,100,AND,20,1,CH,NE,X'40')

Figure 13. Sample OMIT Control Statement

In this example, records will be omitted from the application if the numeric value
in the field beginning in byte 1 is equal to 100 and if the character value in byte 20
is not equal to a blank (X'40').
The next set of control statements exemplifies record selection using bit level logic.
The first two examples involve a comparison between a bit mask (shown coded in
binary and hexadecimal format) and a binary input field. The third example is a
comparison between a bit pattern and a binary field.
Example 3

INCLUDE COND=(10,1,BI,ALL,B'01001000')
or INCLUDE COND=(10,1,BI,ALL,X'48')

Figure 14. Sample INCLUDE Control Statement Using a Bit Mask

The record selection condition has the following elements (from left to right): a
binary field (BI) of length 1 byte that starts at column 10 of the record, a
comparison operator (ALL), and a bit mask (B'01001000' in binary, X'48' in
hexadecimal). Counting from the left, the second and fifth bits of the bit mask are
ON (1). For the selection condition to be true, the same bits must be ON in the
binary input field. Therefore, if the input field contains, for example, 01001000,
01111000 or 11111111, the condition for the inclusion of records is satisfied.
However, if the input field contains a bit string where both mask bits are not ON
(e.g., 01000000, in which the fifth bit is not ON), the condition fails and the records
are omitted.

Syncsort MFX Programmer’s Guide 2–49

INCLUDE/OMIT

Example 4

INCLUDE COND=(10,1,BI,NOTNONE,B'01001000')
or INCLUDE COND=(10,1,BI,NOTNONE,X'48')

Figure 15. Sample INCLUDE Control Statement Using a Bit Mask

The condition for the inclusion of records is met if at least one of the mask bits is
ON in the input field. Therefore, the condition would evaluate as true, if the bit
string in the binary field were 01000000 (the second bit is ON), 000010000 (the
fifth bit is ON), 01001000 (both the second and fifth bit are ON). However, with the
string 10000111, for instance, in the input field, the specified condition would
evaluate as false (resulting in the omission of records), since neither mask bit is
ON.
The above method of comparing a binary input field to a bit mask is useful for
testing the contents of a "flag" byte where each bit has a different meaning.
Example 5

INCLUDE COND=(21,4,BI,EQ,B'000000010001........100100011111')

Figure 16. Sample INCLUDE Control Statement Using a Bit Pattern

The condition specifies a 4-byte long binary input field (BI) in column 21, a logical
relationship (EQ), and a bit pattern. The bit pattern describes the required
sequence of 1s and 0s in the first and last twelve bit positions. The row of periods in
the pattern represents the part of the string that is irrelevant to the definition of
the condition. The condition is true, if the sequence of 1s and 0s in the input field is
identical to that described in the bit pattern.
The method of comparing a binary input field to a bit pattern is useful when testing
for numeric digits that are one half byte each, as in the packed data format. For
example, assume that the binary input field specified in the condition above is a
date field in the PD format X'0mmddyyF'. Each date element is split across a byte
boundary. The second half-byte of each byte (except the last) represents the first of
the two digits that form a date element (mm,dd,yy). (In the last byte, the second
half-byte--1111 in binary and F in hexadecimal--stands for the fact that the bit
pattern encodes a packed decimal.) The first half-byte of each byte (except the first)
represents the second digit of a date element (mm,dd,yy). (The first half-byte, i.e.
0000, of the bit pattern gives it the length specified for the binary field at column
21.) Mapping this scheme onto the bit pattern in the control statement results in
the following.

2–50 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

That is, the above control statement is an instruction to select just those records in
whose date field mm and yy equal 11 and 91, respectively, while dd can have any
value. In other words, the records thus selected are those from November 1991.
Example 6
The following example illustrates substring comparisons.

INCLUDE COND=(11,60,EQ,C'ANYTOWN',
OR,121,3,EQ,C'A01,A05,A06,A09'),FORMAT=SS

Figure 17. Sample INCLUDE Control Statement Using Substring Comparison

In this example, a record will be included in the application if either of the

following conditions is true:
• The literal 'ANYTOWN' is found in the 60-byte field starting at position 11 in
the record.
• The contents of the 3-byte field starting at position 121 matches one of the four
substrings ('A01', 'A05', 'A06', or 'A09') in the constant. Because it is known that
the 3-byte field does not contain any commas, there cannot be a match to the
constants ‘01,’ ‘1,A’, ‘05,’ etc.
Example 7
The following sample control statements illustrate substring comparisons with
various forms of pattern (wildcard) constants.

INCLUDE COND=(20,12,SS,EQ,(C’ST’,*,C’KU’))

Figure 18. Sample INCLUDE Control Statement Using Substring with Wildcard (*)

In this example, a 12-byte field starting in position 20 will be searched for strings
that begin with ST and end with KU anywhere in the field, regardless of the
characters in between. Hence, records with ST43624KU in positions 22 through 30
and ST12KU in columns 24 to 29 would be included, as well as records with STKU
in the field.

Syncsort MFX Programmer’s Guide 2–51

INCLUDE/OMIT

The record selection would be different if the INCLUDE statement were modified
to the following:

INCLUDE COND=(20,12,SS,EQ,(C’ST’,%%,C’KU’))

Figure 19. Sample INCLUDE Control Statement Using Substring with Wildcard (%)

In this case, only the record with ST12KU would be included since only two
characters would be allowed between the ST and the KU character constants.
The record selection would also be different if the INCLUDE statement were
modified to the following:

INCLUDE COND=(20,12,SS,EQ,(C’ST’,*%%%,C’KU’))

Figure 20. Sample INCLUDE Control Statement Using Substring with Wildcard (*%)

In this case, only the record with ST43624KU would be included, since three or
more characters are required between the ST and KU character strings.
Example 8

OMIT COND=(24,3,ZD,EQ,NUM,AND,31,5,ZD,NE,NUM)

OMIT COND=(24,3,EQ,NUM,AND,31,5,NE,NUM),FORMAT=ZD

Figure 21. Sample OMIT Control Statement Using NUM

In this example, both statements are equivalent; the latter statement specifies the
ZD format using the FORMAT=f subparameter. Records will be omitted from the
application if the first field (byte 24 to byte 26) is identified as zoned decimal
numeric AND the second field (byte 31 to byte 35) is identified as zoned decimal
non-numeric.
Example 9
In the following example, two files are similar enough to allow them to be sorted
together using the MULTIIN PARM, but certain fields are in different locations in
the record.

INCLUDE COND=(&MULTIINDD,EQ,C’01’,&,50,20,CH,EQ,C’NEW YORK’,

OR,&MULTIINDD,EQ,C’02’,&,35,18,CH,EQ,C’NEW YORK’)

Figure 22. Sample INCLUDE Control Statement with &MULTIINDD

This statement will include only New York records despite differences in
formatting of the input files defined by SORTMI01 and SORTMI02.

2–52 Syncsort MFX Programmer’s Guide

 INCLUDE/OMIT

Example 10
The following example illustrates an INCLUDE comparison based on CENTWIN
processing.

INCLUDE COND=(20,2,Y2C,GT,96)

Figure 23. Sample INCLUDE Control Statement with CENTWIN Processing

In this example only records whose data are from the years greater than 1996 will
be included in the application. If the CENTWIN parameter were set to 1980,
representing a century window of 1980 to 2079, the records would be processed in
the following manner:

Contents of Record
Positions 20 and 21 Disposition

84 Omitted - represents 1984

99 Included - represents 1999

37 Included - represents 2037

Example 11
The following INCLUDE control statement illustrates the use of the current date
constant and the current date with an offset to include records with dates starting
with the current date and spanning through the two week period prior to the
current date.

INCLUDE COND=(5,8,ZD,LE,&DATE1P,AND,5,8,ZD,GT,&DATE1P-14)

Figure 24. Sample INCLUDE Control Statement Using Current Date Constant and Cur-
rent Date With an Offset Comparison

If the application were run on April 25, 2002, the records included would have
dates in the 8-byte field starting at position 5 from April 12, 2002 through and
including April 25, 2002.
Applications using the INCLUDE/OMIT control statement are illustrated in
“Chapter 3, How to Use the MFX Data Utility Features”.

Syncsort MFX Programmer’s Guide 2–53

INREC

INREC Control Statement

The INREC control statement reformats the input records. Use the INREC control
statement to add, delete, or reformat fields before the records are sorted or merged.
Use the OUTREC control statement or the OUTREC parameter of the OUTFIL
control statement to delete or reformat fields after the records are sorted or
merged. Note that INREC is performed after E15 exit processing and
INCLUDE/OMIT control statement processing.
Using the INREC control statement to delete data fields improves sort performance
by reducing the number of bytes MFX must process. The same result may be
achieved in some cases by changing the data format of certain fields. For example,
if you need to change the format of a ZD field to PD, which reduces the number of
bytes for the field, it is more efficient to use INREC rather than OUTREC for the
conversion. Additionally, for SORT/MERGE processing PD fields are processed
more efficiently than ZD fields.
Except for CONVERT, all the functions performed by the OUTREC control
statement, such as inserting character strings or changing the data format of a
numeric field, can also be performed by the INREC control statement. (See
“OUTREC Control Statement” on page 2-134 for an explanation of these functions.)
For example, you can use the INREC control statement to insert zeros of the proper
format to expand a numeric field before SUM or DUPKEYS processing to prevent
arithmetic overflow. However, you will usually want to use the OUTREC control
statement rather than the INREC control statement to expand the record because
OUTREC processing takes place after records are sorted or merged.
There is one function available with INREC that is not available with the OUTREC
control statement or the OUTREC parameter of the OUTFIL control statement:
the &MULTIINDD subparameter.
If you use the INREC control statement to reformat the input record, remember to
use the post-INREC field positions when you specify the SORT, MERGE, SUM,
DUPKEYS, OUTREC, and/or OUTFIL control statements.
If the SEQNUM function is used in a SORT application to insert a sequence
number field in the record, this field will reflect the order of the records prior to
sorting. In a MERGE application, the field will reflect the order of the records as
they were read from each input in the merge.

2–54 Syncsort MFX Programmer’s Guide

 INREC

INREC Control Statement Format

The format of the INREC control statement is illustrated below:

 
   
  FIELDS  
 [ PARSE=(subparm), ]  BUILD =(fields) 
INREC   OVERLAY  
   
 ( IFTHEN=(subparm) )[,IFTHEN=(subparm), ... ] [ ,IFOUTLEN=n ] 
 FINDREP=(subparm) 
 

where fields are:

OUTREC fields
or &MULTIINDD

Figure 25. INREC Control Statement Format

See “OUTREC Control Statement” on page 2-134 for a complete description of all of
the INREC statement parameters, except for &MULTIINDD.

&MULTIINDD Subparameter (Optional)

The &MULTIINDD subparameter of the INREC FIELDS parameter is used
together with the MULTIIN PARM, which directs MFX to read multiple input files
from separate DD statements. &MULTIINDD is defined as the two-byte C’nn’ of
the SORTMInn ddname (or the two-byte C’n ’ for SORTMIn) from which a record
was read.
&MULTIINDD inserts the two-byte character string identifying the input record’s
origin into the record produced by the INREC statement.
If the MULTIIN PARM option is not in effect, then &MULTIINDD is defined as two
blanks (C’ ‘). If an E15 or E32 exit inserts a record, &MULTIINDD will be C’EX’ for
that record. There will be no change to the &MULTIINDD definition when an E15
exit accepts or changes a record.
&MULTIINDD can only be specified on the INREC statement, including within the
IFTHEN WHEN subparameter on INREC, and not on the OUTREC control
statement or the OUTREC parameter of the OUTFIL control statement.

Syncsort MFX Programmer’s Guide 2–55

INREC

Sample INREC Control Statements

INREC FIELDS=(1:1,20,21:40,15,ZD,PD,29:60,5)

Figure 26. Sample INREC Control Statement

This INREC control statement specifies three data fields from an 80-byte record:
• The first field begins in byte 1 of the input record and is 20 bytes long.
• The second field begins in byte 40 of the input record and is a 15-byte ZD field.
The data format is to be converted to PD. Since the input field contains 15
decimal digits, the converted PD output field created by MFX will be 8 bytes
long.
• The third field begins in byte 60 of the input record and is 5 bytes long.
These three fields have been positioned to begin in bytes 1, 21, and 29, as indicated
by their column prefixes.
The reformatted input record is now just 33 bytes long.

INREC IFTHEN=(WHEN=(&MULTIINDD,EQ,C’01’),BUILD=(C’2007: ‘,1,100)),

IFTHEN=(WHEN=(&MULTIINDD,EQ,C’02’),BUILD=(C’2008: ‘,1,100)),
IFTHEN=(WHEN=(&MULTIINDD,EQ,C’03’),BUILD=(C’2009: ‘,1,100)),

SORT FIELDS=(1,4,CH,A,...)

Figure 27. Sample INREC Control Statement with &MULTIINDD

This INREC control statement can be used when multiple input files are for
different time periods, but do not contain time-distinguishing data.
In this example, a report will be formatted with the year as the primary key when
each of three input files is for a different year.
For comprehensive examples that illustrate the INREC control statement see
“Chapter 3, How to Use the MFX Data Utility Features”.

2–56 Syncsort MFX Programmer’s Guide

 JOIN

JOIN Control Statement

The JOIN control statement specifies the disposition of paired and unpaired
records in a join.
When you do not provide a JOIN control statement in an application that has
JOINKEYS control statements, MFX produces an output from the join operation
that includes all paired records (an “inner join”). All unpaired records from both
SORTJNF1 and SORTJNF2 are discarded. By providing a JOIN control statement,
you can specify that unpaired records are to be included in the join output (an
“outer join”). Parameters of the JOIN statement provide options as to which of the
unpaired records are to be retained for output.
See the descriptions of the JOINKEYS and REFORMAT control statements for
additional information.

JOIN Control Statement Format

The format of the JOIN control statement is illustrated below:

JOIN UNPAIRED

[,F1]

[,F2]

[,ONLY]

Figure 28. JOIN Control Statement Format

Retaining Unpaired Records

When joining files, a record from one file may or may not have a match in the other
file. A match occurs when the contents of the join keys in the record from the first
file equal the contents of the join keys in the record from the second file.
By specifying the JOIN statement you can discard unpaired records from one or
both files, or retain unpaired records from both files.
To retain unpaired records from SORTJNF1 (a “left outer join”) in addition to all
joined records, specify:

JOIN UNPAIRED,F1

Figure 29. Sample JOIN Statement to Retain Unpaired Records from SORTJNF1

Syncsort MFX Programmer’s Guide 2–57

 JOIN

To retain unpaired records from SORTJNF2 (a “right outer join”) in addition to all
joined records, specify:

JOIN UNPAIRED,F2

Figure 30. Sample JOIN Statement to Retain Unpaired Records from SORTJNF2

To retain unpaired records from both SORTJNF1 and SORTJNF2 (a “full outer
join”) in addition to all joined records, specify either:

JOIN UNPAIRED,F1,F2

Figure 31. Sample JOIN Statement to Retain Unpaired Records from SORTJNF1/2

or simply:

JOIN UNPAIRED

Figure 32. Sample JOIN Statement to Retain Unpaired Records from SORTJNF1/2

Discarding Paired Records

You have the option of discarding the paired records from a join and keeping only
the unpaired ones. To do this, specify:

JOIN UNPAIRED,ONLY

Figure 33. Sample JOIN Statement to Discard Paired Records

If you want to keep only the unpaired records from one SORTJNF1 or SORTJNF2,
add either the F1 or the F2 parameter.
Note: See the description of the REFORMAT statement for a discussion on what
will appear in the record created by join processing when source fields from either
SORTJNF1 or SORTJNF2 are not available due to a join unpaired operation.
For more information, see “Joining Records from Multiple Files” on page 3-14.

2–58 Syncsort MFX Programmer’s Guide

 JOINKEYS

JOINKEYS Control Statement

Use the JOINKEYS statement to enable join feature processing and to identify the
fields used to select records for join processing.
The join feature joins records from two input files that are specified on the
SORTJNF1 and SORTJNF2 DD statements. By default, when the JOINKEYS
fields from m records in SORTJNF1 match the JOINKEYS fields from n records in
SORTJNF2, all combinations of the records are joined using the REFORMAT
statement, producing m*n records as input to subsequent MFX processing. (This is
called an “inner join.”)
See the discussion of the REFORMAT control statement for a description of how a
record is constructed from the two records that have been selected as a match.
If the optional JOIN UNPAIRED statement is specified, the unmatched records
from the SORTJNF1 and/or SORTJNF2 files will also be REFORMATted and
included in the input to MFX without being joined. (Including the unmatched
records from SORTJNF1 is called a “left outer join,” including the unmatched
records from SORTJNF2 is called a “right outer join,” and including all unmatched
records is called a “full outer join.”) Optionally, only these unmatched records will
become input to MFX. See the descriptions of the JOIN and REFORMAT
statements for further details on their specification.
The input files do not need to be presorted or have the same record type.
Two JOINKEYS control statements are required – one for each of the two files used
in the join.
The JOINKEYS control statement cannot be used with MAXSORT, PARASORT,
Syncsort PipeSort, SKIPREC, checkpoint, and merge exits (except for E35), and the
DB2 Query feature.

Syncsort MFX Programmer’s Guide 2–59

JOINKEYS

JOINKEYS Control Statement Format

The format of the JOINKEYS control statement is illustrated below:

 
 FILE= F1 
 F2 ,FIELDS =(p ,l [ ,f ],o [,p ,l [ ,f ],o ]...)[,FORMAT=f]
JOINKEYS  1 1 1 1 2 2 2 2
 F1=ddname 
 F2=ddname 
 
[,SORTED][,NOSEQCK]

 ALL 
 NONE 
 
 ,INCLUDE    ,AND,  
 ,OMIT =  ,&,  
   (c   c 2 ... ) 
 1 
 ,OR, 
  ,|,  
 

,TASKID=xx

F 
,TYPE=  
V 

,STOPAFT=n

Figure 34. JOINKEYS Control Statement Format

FILE Parameter (Required)

The FILE parameter connects the JOINKEYS control statement with the input file
to be read. The specification of F1 connects the JOINKEYS control statement with
the SORTJNF1 DD statement. The specification of F2 connects the JOINKEYS
control statement with the SORTJNF2 DD statement. FILE cannot be used if
either F1 or F2 is used.
For large applications, if one of the two input files has many more duplicate keys
for the join than the other input file, that file should be allocated as SORTJNF2 to
achieve optimal performance.

2–60 Syncsort MFX Programmer’s Guide

 JOINKEYS

The format of the FILE parameter is illustrated in the following figure.

 F1 
FILE=  
 F2 

Figure 35. FILE Parameter Format

F1 and F2 Parameters (Required)

The F1=ddname and F2=ddname parameters are used to specify the ddnames of
the input files to be read for the JOINKEYS control statement. F1 is used in place
of FILE=F1 to connect JOINKEYS to the first input file and change the ddname
from SORTJNF1. F2 is used similarly to connect JOINKEYS to te second input file
and change the ddname from SORTJNF2. You should use only one of F1 or F2. The
FILE parameter cannot be used if either F1 or F2 has been used.
The format of the F1 and F2 parameters is illustrated below

 F1=ddname 
 
 F2=ddname 

Figure 36. F1 and F2 Parameter Format

FIELDS Parameter (Required)

The FIELDS parameter is required. It describes the fields to be used to match
records from the two files, SORTJNF1 and SORTJNF2.
The number of JOINKEYS fields and their sorted order (A or D) must be the same
for both files, although their starting positions and lengths need not be the same.
The join files do not need to be presorted on the fields specified on the JOINKEYS
statement. By default, MFX will sort the records to the proper sequence before
performing the join operation. If one or both of the files are already in the
JOINKEYS fields sequence, the SORTED parameter (see below) of the JOINKEYS
statement can be specified. If the SORTED parameter can be used, the
performance of the application will be improved since the need for MFX to preorder
the records prior to join processing will be removed.
The maximum number of JOINKEYS fields is 64.
Each JOINKEYS field may be anywhere within the record through column 32750,
the maximum length of a field is 4080 bytes, and the sum of all fields on a
JOINKEYS statement cannot exceed 4080 bytes.
Each field specified in the FIELDS parameter is identified by a position (p), length
(l), format (f), and order (o).

Syncsort MFX Programmer’s Guide 2–61

JOINKEYS

p The position value indicates the first byte of the field relative to the
beginning of the input record.
l The length value indicates the length of the control field.
f The format value indicates the data format. For a list of valid formats,
refer to the table in the next section, “Valid Formats for JOINKEYS
Fields.” If all the fields have the same format, you can specify the format
value once by using the FORMAT=f subparameter. If you specify both
the individual f values and the FORMAT subparameter, the individual f
values will be used for fields where they are specified.
If the format value is omitted, BI (binary) format will be assumed.
o The order value indicates the collating sequence of the field:
A=Ascending order
D=Descending order
Valid Formats for JOINKEYS Fields
Table 11 on page 2-62 lists the valid formats for JOINKEYS fields.

Acceptable Field
Data Format
Length (Bytes)

AQ 1 to 4080

BI* 1 to 4080

CH 1 to 4080

FI 1 to 256

PD 1 to 255

ZD 1 to 256

Note: *Bit fields are not permitted.

Table 11. Valid Formats and Lengths of JOINKEYS Fields

Note that LOCALE will not be used with CH fields.

Field-to-Field Comparisons
The formats of the JOINKEYS fields for the SORTJNF1 file must be compatible
with the corresponding fields for the SORTJNF2 file.
Table 12 on page 2-63 shows the permissible types of format comparisons.

2–62 Syncsort MFX Programmer’s Guide

 JOINKEYS

AQ BI CH FI PD ZD

AQ X

BI X X

CH X X

FI X

PD X X

ZD X X

Table 12. Permissible Field-to-Field Comparisons for JOINKEYS Formats

Padding of Compared Fields of Unequal Lengths

When two fields of unequal lengths are compared, the shorter field is padded to the
length of the longer field. Padding takes place as follows:
• The padding characters are blanks when the shorter field is in CH or AQ
format; otherwise, they are zeros of the shorter field’s own format. For negative
FI fields, the padding character is X‘FF’.
• Padding is on the right if the shorter field is in BI, CH or AQ format. Padding is
on the left for FI, PD and ZD formats.

SORTED Parameter (Optional)

By default, MFX will presume that the records in the file are not presequenced per
the JOINKEYS fields specified. If the records are already collated in the proper
sequence, the SORTED parameter can be specified to improve the application's
performance. Since LOCALE is not used for CH JOINKEYS fields, do not specify
SORTED if the files were previously sorted with LOCALE.
MFX will sequence check each input file according to its JOINKEYS fields. If the
SORTED parameter of the JOINKEYS statement was specified to indicate that the
file was presorted and the sequence check fails, MFX will issue a critical error
message containing the file number. The record number within the file will also be
in the error message text whenever the INCLUDE/OMIT parameter of the
JOINKEYS statement was not specified.

NOSEQCK Parameter (Optional)

The NOSEQCK parameter may be used when the SORTED parameter has been
specified. NOSEQCK instructs MFX to bypass the sequence check that MFX
performs for the sorted input file. NOSEQCK should only be used when you are
certain that the input file connected to the JOINKEYS statement has already been
sorted in the same collating sequence as specified in the JOINKEYS FIELDS
parameter, otherwise your output may be incorrect. NOSEQCK may slightly

Syncsort MFX Programmer’s Guide 2–63

JOINKEYS

improve the performance of your JOINKEYS application. NOSEQCK is ignored if

SORTED has not been specified.

INCLUDE/OMIT Parameter (Optional)

Specify the INCLUDE or OMIT parameter to indicate which records are to be
included or omitted from the SORTJNFn file specified on the JOINKEYS
statement. The INCLUDE/OMIT processing occurs prior to the JOINKEYS field
matching process.
The format for the INCLUDE/OMIT parameter is illustrated below:

 ALL 
 NONE 
 
 INCLUDE    ,AND,  
 =   
 OMIT   (c 
,&,
 c 2 ... ) 
 1 
 ,OR, 
  ,|,  
 

Figure 37. INCLUDE/OMIT Parameter Format

See “INCLUDE/OMIT Control Statement” on page 2-29 for the detailed format of a
comparison. The FORMAT=f parameter, which is permitted for the
INCLUDE/OMIT control statement, is not permitted for the INCLUDE/OMIT
parameter. Field formats must be specified on a field-by-field basis.

TASKID Parameter (Optional)

The TASKID=xx parameter is used to change the first two bytes of the ddnames for
the dynamically allocated sortwork data sets used to sort the JOINKEYS input file.
This parameter should be used when invoking MFX from a program and attaching
multiple join applications that will run concurrently.

TYPE Parameter (Optional)

The TYPE parameter can be used to indicate the record format. TYPE=F indicates
fixed-length records; TYPE=V indicates variable-length records.
TYPE should be provided if the input file being specified is VSAM. If TYPE is not
provided, TYPE=F will be assumed if the SORTJNFn file is VSAM.
Note: If the TYPE specification differs from the RECFM DCB parameter for the
SORTJNFn DD statement, the latter takes precedence.
For more information, see “Joining Records from Multiple Files” on page 3-14.

2–64 Syncsort MFX Programmer’s Guide

 JOINKEYS

STOPAFT Parameter (Optional)

The STOPAFT parameter limits the number of records processed from the
SORTJNFn file specified. This can be useful when testing a new join application.
You can sample a subset of one or both input files and view your output without
having to sort both input files in their entirety and possibly generate a very large
number of joined records.
The variable n specifies the number of records to be sorted or copied from
SORTJNF1 or SORTJNF2. These will be the first n records after JOINKEYS
INCLUDE/OMIT processing, if specified, has completed.

Syncsort MFX Programmer’s Guide 2–65

MERGE

MERGE Control Statement

The MERGE control statement is required for every merge application. The
MERGE control statement can also define a copy application.

Cultural Environment Support

Cultural environment support allows you to choose an alternative set of collating
rules based on a specified national language. The alternative collating applies to
SORT/MERGE and INCLUDE/OMIT processing.
For additional detail, see “LOCALE” on page 5-18.

MERGE Control Statement Format

The format of the MERGE control statement is illustrated below:

FIELDS Parameter (Required for a Merge)

MERGE  FIELDS =(p 1 ,l 1 [ ,f 1 ],o 1 [,p 2 ,l 2 [ ,f 2 ],o 2 ]...)[,FORMAT=f] 

 
 FIELDS = COPY 

0 
  ,CKPT ,EQUALS
,CENTWIN=  s  ,NOEQUALS [ ,FILES=n ]
f ,CHKPT
 
[ ,SKIPREC=n ] [ ,STOPAFT=n ]

Figure 38. MERGE Control Statement Format

The FIELDS parameter is required for a merge. It describes the control fields.
List the control fields in order of greatest to least priority, with the primary control
field listed first, followed by progressively less significant fields. You can specify up
to 128 control fields; however, if fields require complex internal processing, the
limit for a particular execution may be less than 128.
Each field specified in the FIELDS parameter is identified by its position p, length
l, format f and order o.
p The position value indicates the first byte of the field relative to the
beginning of the input record after INREC and/or E32 processing, if
specified, have completed.
Binary control fields can begin on any bit of a byte. When a binary field
does not begin on a byte boundary, you must specify the bit number (0-
7). For example, a position value of 21.3 refers to the 4th bit of the 21st
byte of the record.

2–66 Syncsort MFX Programmer’s Guide

 MERGE

l The length value indicates the length of the control field. The length
value must be an integer number of bytes, except for the length of a
binary control field which can be specified in bits. For example, a length
value of 0.5 refers to a binary control field 5 bits long.
For signed fields, the length value must include the area occupied by the
sign.
f The format value indicates the data format. For a list of valid formats,
refer to Table 13 on page 2-67. If all the control fields have the same for-
mat, you can specify the format value once by using the FORMAT=f sub-
parameter. If you specify both the individual f values and the FORMAT
subparameter, the individual f values will be used for fields where they
are specified.
o The order value indicates how the field is to be collated:
A=Ascending order
D=Descending order
E=As modified by an E61 exit

Valid Formats for Merge Control Fields

The following table lists the valid formats for merge control fields.

Field Length
Code Data Format
(bytes)

AC EBCDIC characters are translated to their ASCII equivalents before 1 to 4091†

merging.

AQ Character. Records are merged according to an alternate sequence spec- 1 to 4091†

ified either in the ALTSEQ control statement or as an installation
default.

ASL Leading separate sign. An ASCII + or - precedes numeric field. One digit 2 to 256
per byte.

AST Trailing separate sign. An ASCII + or - trails numeric field. One digit 2 to 256
per byte.

BI Binary. Unsigned. 1 bit to 4092*

CH Character. Unsigned. 1 to 4092*

CLO Leading overpunch sign. Hexadecimal F,C,E, or A in the first 4 bits of 1 to 256
OL your field indicates a positive number. Hexadecimal D or B in the first 4
bits indicates a negative number. One digit per byte. CMP=CLC is
forced.

Table 13. (Page 1 of 3)Format Code Chart

Syncsort MFX Programmer’s Guide 2–67

MERGE

Field Length
Code Data Format
(bytes)

CSF Floating sign format. An optional leading sign may be specified immedi- 1 to 32
FS ately to the left of the digits. If the sign is a -, the number is treated as
negative. For other characters, the number is treated as positive. Char-
acters to the left of the sign are ignored.

CSL Leading separate sign. An EBCDIC + or - precedes numeric field. One 2 to 256
LS digit per byte. CMP=CLC is forced.

CST Trailing separate sign. An EBCDIC + or - follows numeric field. One 2 to 256
TS digit per byte. CMP=CLC is forced.

FD Decimal floating point. Signed. An SNaN or QNaN value is invalid and 4, 8, or 16

will cause a WER497A error.

FI Fixed point. Signed. (Equivalent to Signed Binary.) 1 to 256

FL Floating point. Normalized. Signed. 2 to 256

PD Packed decimal. Signed. 1 to 256

PD0 Packed decimal. 2-8-byte packed decimal data with the first digit and 2-8
trailing sign ignored. The remaining bytes are treated as packed deci-
mal digits. Typically PD0 is used with century window processing and
Y2P format; Y2P processes the year, while PD0 processes month and
day.

SFF Signed free format. Decimal digits (0-9) are extracted from right to left 1 to 44
to form a number value. A character of – or ) found within the field will
cause the value to be treated as a negative number. All other non-deci-
mal digit values in the field are ignored. A maximum of 31 digits can be
provided. When more than 31 digits are found in the field, the leftmost
digits will be ignored.

UFF Unsigned free format. Decimal digits (0-9) are extracted from right to 1 to 44
left to form a positive number value. All non-decimal digit values in the
field are ignored. A maximum of 31 digits can be provided. When more
than 31 digits are found in the field, the leftmost digits will be ignored.

Y2B Binary. 2-digit, 1-byte binary year data treated as a 4-digit year by 1
CENTWIN (century window) processing.

Y2C Character. 2-digit character year data treated as a 4-digit year by 2

CENTWIN (century window) processing. Processing is identical to Y2Z
fields.

Y2D Packed decimal. 2-digit, 1-byte packed decimal year data treated as a 4- 1
digit year by CENTWIN (century window) processing.

Table 13. (Page 2 of 3)Format Code Chart

2–68 Syncsort MFX Programmer’s Guide

 MERGE

Field Length
Code Data Format
(bytes)

Y2P Packed decimal. 2-digit, 2-byte packed decimal year data. Of the four 2
packed digits contained in the 2 bytes, the first digit and trailing sign
are ignored; the two inner digits are treated as a 4-digit year by
CENTWIN processing.

Y2S Character or zoned decimal. 2-digit, 2-byte valid numeric data treated 2
as a 4-digit year by CENTWIN (century window) processing, as for Y2C
and Y2Z. However, certain data are not treated as year data. Data with
binary zeros (X'00') or a blank (X'40') in the first byte will be collated
before valid numeric year data for ascending order (after year data for
descending order). Data with all binary ones (X'FF') in the first byte will
be collated after valid numeric year data for ascending order (before
year data for descending order). Zones are ignored, as for Y2C and Y2Z,
except for data where the first byte begins with X'00', X'40' or X'FF'.

Y2T Full-date, character, binary, or packed decimal formats. Full-date data 2-6
formats can be used to merge a variety of date fields. They can process
Y2U dates ending or starting with year digits (x...xyy or yyx...x). They can
also process non-date data commonly used with dates. For details, see
Y2V
Table 39 on page 239.
Y2W

Y2X

Y2Y

Y2Z Zoned decimal. 2-digit, 2-byte zoned decimal year data treated as a 2
4-digit year by CENTWIN (century window) processing. The zones are
ignored. Processing is identical to Y2C fields.

ZD Zoned decimal. Trailing overpunch in the first 4 bits of the rightmost 1 to 256
CTO byte gives the sign. Hexadecimal F,C,E, or A indicates a positive num-
OT ber. Hexadecimal D or B indicates a negative number. One digit per
byte. CTO forces CMP=CLC.

Note: * 4084 for variable-length records.

† 2043 for variable-length records.

Table 13. (Page 3 of 3)Format Code Chart

For information on the year data formats (Y2B, Y2C, Y2D, Y2P, Y2S and Y2Z) plus
the related data format PD0, see “CENTWIN Parameter (Optional)” on page 2-71
and “Converting Year Data with Century Window Processing on INREC, OUTREC,
or OUTFIL OUTREC” on page 2-162. Also, see “Specifying Field-to-Field Standard
Comparisons for Year Fields” on page 2-39.

Syncsort MFX Programmer’s Guide 2–69

MERGE

Rules for Specifying Merge Control Fields

• For fixed-length records, the sum of the lengths of all control fields cannot
exceed 32752 bytes. When EQUALS is in effect, the sum of their lengths cannot
exceed 4088 bytes.
• For variable-length records, all control fields must be located within the first
32750 bytes and the sum of their lengths cannot exceed 4084 bytes. When
EQUALS is in effect, all control fields must be located within the first 32746
bytes and the sum of their lengths cannot exceed 4080 bytes.
• Control fields can be in contiguous or non-contiguous locations in the record.
• Remember that for variable-length records, the first 4 bytes are reserved for the
Record Descriptor Word, so the first byte of the data portion of the record is byte
5.
• If the output file is a key-sequenced VSAM cluster, the VSAM key must be the
first control field specified.

Comparing PD and ZD Control Fields

The CMP PARM determines how PD and ZD control fields will be compared. When
CMP=CPD is in effect, the Compare Decimal (CP) instruction may be used under
certain circumstances for the compare. ZD fields are packed and then compared.
This method has performance advantages. However, invalid PD data may cause a
system 0C7 abend and program termination. Moreover, the integrity of ZD fields is
only guaranteed when they contain valid ZD data. The CMP=CPD method will not
be used for control fields that exceed 16 bytes or for variable-length merges when
an even value (0, 2, 4, or 6) is specified for the VLTEST PARM.
When CMP=CLC is in effect, no data validation is performed and the integrity of
the output is maintained, even if the sign for a PD or ZD field is invalid. This
method will be used if any control field exceeds 16 bytes or for variable-length
merges when an even value is specified for the VLTEST PARM.

FIELDS=COPY (Required for a Copy)

Use FIELDS=COPY to copy one or more input files. (Multiple files can be copied if
they are concatenated on the SORTIN DD specification.) Other control statements
such as INCLUDE/OMIT, INREC, OUTREC, and OUTFIL may be specified in
conjunction with a copy application, allowing you to edit and reformat the file(s)
without any collation processing.
The SUM or DUPKEYS control statement and an E32 exit should not be specified
with FIELDS=COPY. All Phase 3 exits can be used.
The SORTIN DD statement defines the input to be copied. (SORTINnn DD
statements are not processed when FIELDS=COPY is specified.)

2–70 Syncsort MFX Programmer’s Guide

 MERGE

CENTWIN Parameter (Optional)

The CENTWIN run-time or installation option acts on 2-digit year data. At run-
time, CENTWIN can be specified as either a PARM option or a SORT/MERGE
control statement parameter. CENTWIN generates a century window (for example,
1950 through 2049) that determines the century to which a 2-digit year belongs.
CENTWIN ensures that year data spanning centuries will be sequenced correctly.
Without CENTWIN processing, an ascending collation would sequence the year 01
before the year 98. With CENTWIN processing, the 01 field could be recognized as
a twenty-first century date (2001) and would thus be sequenced after 98 (1998).
For more information on specifying the CENTWIN option, see “CENTWIN” on
page 5-6.
CENTWIN processing only applies to data defined as year data formats (Y2B, Y2C,
Y2D, Y2P, Y2S, and Y2Z) and the full-date formats (Y2T, Y2U, Y2V, Y2W, Y2X, and
Y2Y). These data formats enable MFX to process 2-digit year fields as 4-digit years.
A related data format, PD0, can be used to process the month and day portions of
packed decimal date fields. To correctly specify date fields for CENTWIN MERGE
processing, you should be familiar with the CENTWIN-related data formats.
The following describes each of the year data formats and provides MERGE control
statement examples:
The Y2B Format
This format is used to sequence 2-digit, 1-byte binary year data with CENTWIN
processing. The binary values are converted to decimal, and the two low order
digits are used as year data. Thus, while binary and decimal values range from 00
to 255, year values range from 00 to 99. The relationship between binary, decimal
and year values is shown in the following table:

Binary Value Decimal Value Year Value

X'00' to X'63' 00 to 99 00-99

X'64' to X'C7' 100 to 199 00-99

X'C8' to X'FF' 200 to 255 00-55

Table 14. Possible Values Representing Year Data with Y2B

The Y2C and Y2Z Formats

These formats represent 2-digit, 2-byte year data in either character (Y2C) or
zoned decimal (Y2Z) format. Either Y2C and Y2Z formats can be used with data of
the form
X'xyxy'
where y is a hexadecimal year digit 0-9 and x is hexadecimal 0 through F. Y2C and
Y2Z ignore the x digits, leaving yy, the 2-digit unsigned year representation.

Syncsort MFX Programmer’s Guide 2–71

MERGE

Suppose you have a character or zoned decimal date field mmddyy that begins at
byte 20. You can use either Y2C or Y2Z to process the yy field. As the following
example indicates, you could specify three merge keys to correctly process this date:

MERGE FIELDS=(24,2,Y2C,A, * Collates yy field as 4-digit year

20,2,CH,A, * Collates mm field
22,2,CH,A) * Collates dd field

Figure 39. Sample MERGE Statement

The yy field (24,2) will be processed according to the century window setting. For
example, if CENTWIN=1945, the field yy=45 will be sequenced as if it were 1945,
and yy=44 would be sequenced as if it were 2044. Thus, for an ascending sequence,
44 would follow 45.
The Y2D Format
This format is used to sequence 2-digit, 1-byte packed decimal year data with
CENTWIN processing. Use Y2D to extract the year data yy from packed decimal
date fields. For example, consider a 3-byte packed decimal data field defined as
X'yyddds'
This field has the year yy in the first byte and the day ddd in bytes 2 and 3. The
packed decimal sign s would be in the last digit (half byte) of the third byte. To
merge this date field, which begins at byte 20, with 4-digit year processing, use the
following MERGE control statement:

MERGE FIELDS=(20,1,Y2D,A, * Collates 2-digit year as 4-digit year

21,2,PD,A) * Collates ddds as 3 digits (ddd)

Figure 40. Sample MERGE Statement

The Y2P Format

This format is used to sequence 2-digit, 2-byte packed decimal year data with
CENTWIN processing. Use Y2P to extract the year data yy from packed decimal
date fields spanning 2 bytes. For example, a packed decimal date of the form
yymmdd would be stored as 4 bytes:
yymmdd = X'0yymmddC'
where the trailing C (sometimes F) is a positive sign and the leading 0 pads the
field on the left to make an even number of digits.
Notice that the components of the date span bytes:
0y ym md dC

2–72 Syncsort MFX Programmer’s Guide

 MERGE

Y2P handles this condition by ignoring the first and last half bytes of the 2-byte
field specification. Thus, Y2P processes 0yym as yy, ignoring the leading digit (0)
and the trailing digit m that is part of the month.
The following example uses Y2P to collate the year portion of the date field, which
begins at byte 20:

MERGE FIELDS=(20,2,Y2P,A) * Collates yy field as 4-digit year

Figure 41. Sample MERGE Statement

The field specification 20,2,Y2P treats X'0yym' as X'yy', and CENTWIN processing
merges yy as a 4-digit year yyyy.
The PD0 format, described below, can assist Y2P by processing month and day data
that overlap year data in the original field.
The Y2S Format
This format is used to sequence 2-digit, 2-byte character or zoned decimal data. The
Y2S format is identical to Y2C and Y2Z for valid numeric data, but Y2S treats data
that begin with X'00', X'40' or X'FF' as non-year data. Thus, the Y2S format can
distinguish records that have non-year data in the first byte of the year field,
allowing such records to be collated differently from other records.
Y2S treats non-year data as follows:
• Data with binary zeros (X'00') or a blank (X'40') in the first byte will not have
century window processing applied to it. Instead, such data will be collated in
sequence, before valid numeric year data for ascending order or after the year
data for descending order.
• Data with all binary ones (X'FF') in the first byte will also not have century
window processing applied to it. Instead, such data will be collated after valid
year numeric data for ascending order or before the year data for descending
order.
Zones are ignored, as for Y2C and Y2Z, except for data where the first byte begins
with X'00', X'40' or X'FF'.
As an example, suppose you want to preserve the input order of header and trailer
records at the start or end of the file, and your header/trailer records are identified
by binary zeros (X'00'), a blank (X'40') or binary ones (X'FF') in the first byte of the
date field. The Y2S format allows CENTWIN to identify the header/trailer records
and treat them differently from other records.
The PD0 Format
This format is used to sequence 2-8 byte packed decimal data. PD0 ignores the first
digit and trailing sign during processing. PD0 is normally used in conjunction with
the Y2P data format. The Y2P format is used to process the 2-digit year portion of a

Syncsort MFX Programmer’s Guide 2–73

MERGE

packed decimal date field, while the PD0 format is used to process the month and
day portion of the field.
Although PD0 is typically used with Y2P, CENTWIN processing is not applied to
PD0.
Consider the packed decimal date field used in the Y2P example above:
yymmdd = X'0yymmddC'
where the trailing C (sometimes F) is a positive sign and the leading 0 pads the
field on the left to make an even number of digits.
Notice that the components of the date span bytes:
0y ym md dC
The date can be processed as follows:
• Y2P processes the year component X'0yym' as X'yy'.
• PD0 processes the month and day components X'ymmddC' as X'mmdd'.
The following MERGE control statement can be used to collate the entire date with
CENTWIN processing:

MERGE FIELDS=(20,2,Y2P,A, * Treats X'0yym' as X'yy'; collates yy as yyyy

21,3,PD0,A) * Treats X'ymmddC' as X'mmdd'

Figure 42. Sample MERGE Statement

Full-Date Formats
Full-date formats can be used to merge various date fields, processing dates ending
or starting with year digits. They also process non-date data that are used with
dates. For a full description of full-date formats, see the following section.

Using Full-Date Formats with CENTWIN

MFX’s full-date data formats enable you to merge a variety of date fields. The full-
date formats are Y2T, Y2U, Y2V, Y2W, Y2X, and Y2Y. These date formats can
process dates ending or starting with year digits:
• x...xyy (for example: qyy, mmyy, dddyy, or mmddyy)
• yyx...x (for example: yyq, yymm, yyddd, or yymmdd)
The full-date formats also process non-date data commonly used with the dates.
MFX interprets two-digit years (yy) according to the century window specified by
the CENTWIN option. CENTWIN processing does not apply to non-date data.
In most cases, for CH, ZD, and PD date fields the full-date data formats are easier
to use than the 2-digit date formats. The 2-digit formats can be more difficult
because you must divide the date into its components. This requires care,
particularly for PD dates, where date components (q, dd, mm, or yy) may span

2–74 Syncsort MFX Programmer’s Guide

 MERGE

bytes or occupy only part of a byte. The full-date formats, on the other hand,
process such dates automatically.
Table on page 2-76 describes the full-date formats. For date forms not in the table,
use the 2-digit year formats or the non-year formats.
Note the following symbols used in Table 15:
y year digit (0-9)
x non-year digit (0-9)
s sign (hexadecimal A-F)
0 unused digit

Full-Date Data Example Date

Date Form Length (bytes)
Format Format Form

Y2T CH, BI yyx yyq 3

yyxx yymm 4

yyxxx yyddd 5

yyxxxx yymmdd 6

Y2U PD yyx yyq 2

(X'yyxs')

yyxxx yyddd 3

(X'yyxxxs')

Y2V PD yyxx yymm 3

(X'0yyxxs')

yyxxxx yymmdd 4
(X'0yyxxxxs')

Y2W CH, BI xyy qyy 3

xxyy mmyy 4

xxxyy dddyy 5

xxxxyy mmddyy 6

Y2X PD xyy qyy 2

(X'xyys')

xxxyy dddyy 3

(X'xxxyys')

Table 15. Full-Date Formats

Syncsort MFX Programmer’s Guide 2–75

MERGE

Full-Date Data Example Date

Date Form Length (bytes)
Format Format Form

Y2Y PD xxyy mmyy 3

(X'0xxyys')

xxxxyy mmddyy 4
(X'0xxxxyys')

Table 15. Full-Date Formats

Table 15 indicates the full-date formats that can be used with character (CH),
binary (BI), or packed decimal (PD) data. Note the recognized non-date values:
Character or binary (Y2T and Y2W full-date formats)
C'0...0' (CH zeros)
C'9...9' (CH nines)
Z'0...0' (ZD zeros)
Z'9...9' (ZD nines)
X'00...00' (BI zeros)
X'40...40' (blanks)
X'FF...FF' (BI ones)
Packed (Y2U, Y2V, Y2X, and Y2Y full-date formats)
P'0...0' (PD zeros)
P'9...9' (PD nines)
The following two examples illustrate how you might use Table (“Table 15
indicates the full-date formats that can be used with character (CH), binary (BI), or
packed decimal (PD) data. Note the recognized non-date values:”) on page 2-76:
• Suppose you have a packed decimal (PD) date field of the form mmyy. To merge
this field correctly, you would use the Y2Y 3-byte format from the table. Thus, if
the field starts in position 30 and the records are in descending order, you would
specify the following MERGE control statement:
MERGE FIELDS=(30,3,Y2Y,D)
Any PD fields of all PD zeros or all PD nines will be processed automatically as
non-date data.
• Suppose you have a character (CH) date field of the form yymmdd. To merge
this field correctly, you would use the Y2T 6-byte format from the table. Thus, if
the field starts in byte 40 and the records are in ascending order, you would
specify the following MERGE control statement:
MERGE FIELDS=(40,6,Y2T,A)

2–76 Syncsort MFX Programmer’s Guide

 MERGE

Any CH zeros, CH nines, BI zeros, blanks, and BI ones will be processed

automatically as non-date data.
Collating Sequence with Full-Date Formats
For full-date formats, the yy component is always processed first (treated as
primary key). This is so even when the yy is physically at the rightmost end of the
field, as for Y2W, Y2X, and Y2Y. For example, a 6-byte Y2W field has the form
xxxxyy. This is collated with the yy as the primary key and xxxx as the secondary
key. Because MFX automatically collates the year character first, you don’t have to
deal with yy manually, for example by using PD0 and Y2D.
It is important to understand that the xxxx component of a full-date format must
be designed to collate as a unit. Suppose you have the 6-byte Y2T field yyxxxx. If
you collate this field in ascending order, then yy collates first (the primary key)
with xxxx collating second (secondary key). Consider two possibilities:
• If yyxxxx is actually yymmdd, you will be merging first by year, then month,
then day.
• If yyxxxx is actually yyddmm, you will merging by year, then day, then month.
In most cases, collating in this way would not be what you intended.
To correctly collate a date, the date components must be in an order suitable for
collating. For example, mmddyy and yymmdd will collate correctly, but ddmmyy or
yyddmm will not. For date forms that will not collate correctly, you must use one of
the 2-digit year formats (Y2B, Y2C, Y2D, Y2P, Y2S, and Y2Z).
The following table shows the order for ascending collation when using full-date
formats with the CENTWIN option:

Full-Date Format Date Format Ascending Sequence

Y2T CH, BI BI zeros

Y2W Blanks

CH/ZD zeros

Lower century dates (e.g. 1980)

Higher century dates (e.g. 2010)

CH/ZD nines

BI ones

Y2U PD PD zeros

Y2V Lower century dates (e.g. 1980)

Y2X Higher century dates (e.g. 2010)

Y2Y PD nines

Table 16. Ascending Sequences

Syncsort MFX Programmer’s Guide 2–77

MERGE

For descending sequence, the collation order is reversed.

Other date formats (non-full-date), with the exception of Y2S, do not process
non-date data; their sequence for ascending order begins with lower century
dates and ends with higher century dates.

CKPT/CHKPT Parameter (Optional)

The CKPT/CHKPT parameter instructs MFX to take a checkpoint at every
end-of-volume of a SORTOUT data set when OUTFIL is not used. Either spelling is
accepted.
This parameter requires a SORTCKPT DD statement. It cannot be specified in
conjunction with a user-issued STIMER macro or an incore sort. Checkpoints
cannot be taken within a user exit routine.
See “Chapter 14 Performance Considerations” for an explanation of the
Checkpoint/Restart feature.

EQUALS/NOEQUALS Parameter (Optional)

The EQUALS parameter insures that equal-keyed records are merged in the order
of their respective files. Equal-keyed records from the lowest numbered SORTINnn
file are written before those from the second input file, etc. NOEQUALS, the
default, specifies that equal-keyed records from different files be written in random
order.
The order of equal-keyed records within each input file is always preserved during
a merge, whether or not the EQUALS parameter is specified.
When the EQUALS parameter is used with the SUM or DUPKEYS control
statement, the first of the equal-keyed records is retained with the sum or
DUPKEYS function value; all other records are deleted after the specified field(s)
have been calculated.
EQUALS/NOEQUALS can also be specified as a PARM option on the EXEC
statement. If this option is specified both on the MERGE control statement and as
a PARM option, the MERGE specification takes precedence.

FILES Parameter (Optional)

The FILES=n parameter specifies the number of input files that an E32 exit will
supply to the merge. The n value can be any number up to 100.
Specifying the FILES parameter both on the MERGE control statement and in the
24-bit parameter list will cause MFX to terminate with a critical error.
The FILES parameter cannot be specified as a PARM on the EXEC statement or in
a $ORTPARM data set.

2–78 Syncsort MFX Programmer’s Guide

 MERGE

SKIPREC Parameter (Optional)

The SKIPREC=n parameter instructs MFX to skip a decimal number of records
before the input file is copied. The n records skipped are deleted from the input file
before INCLUDE/OMIT processing, if specified, takes place.
The SKIPREC parameter should only be specified for a MERGE FIELDS=COPY
operation. SKIPREC will be ignored when doing a merge of SORTINnn data sets.
If SKIPREC is specified as a PARM option as well as on the MERGE control
statement, the PARM specification takes precedence.

STOPAFT Parameter (Optional)

The STOPAFT=n parameter specifies the number of records to be copied. These
will be the first n records after INCLUDE/OMIT and SKIPREC processing, if
specified, have completed.
The STOPAFT parameter should only be specified for a MERGE FIELDS=COPY
operation. STOPAFT will be ignored when doing a merge of multiple SORTINnn
data sets.
If STOPAFT is specified as a PARM option as well as on the MERGE control
statement, the PARM specification takes precedence.

Sample MERGE Control Statements

MERGE FIELDS=(1,5,CH,A,10,2,PD,D,30,4,BI,A)

Figure 43. Sample MERGE Control Statement

This sample MERGE control statement specifies three merge control fields:
• The first, or primary, control field begins in byte 1, is 5 bytes long, is in
character format and is to be merged in ascending order.
• The second control field begins in byte 10, is 2 bytes long, is in packed decimal
format and is to be merged in descending order.
• The third control field begins in the third bit of byte 30, is 4 bytes long, is in
binary format and is to be merged in ascending order.

MERGE FIELDS=COPY,STOPAFT=200

Figure 44. Sample MERGE Control Statement

This MERGE statement specifies a copy operation. Only the first 200 records will
be copied.

Syncsort MFX Programmer’s Guide 2–79

MODS

MODS Control Statement

The MODS control statement specifies a user exit routine and is required with an
exit. Refer to “Chapter 17 Messages” for a detailed explanation of how to specify
exit programs.

MODS Control Statement Format

The format of the MODS control statement is illustrated below.

MODS exit-name1=(parameters1),...,exit-name16=(parameters16)

where parameters =

 ,N 
 ,S 
 ,C 
 
r,b [,d]  ,E 
 ,X 
 ,T 
 
 ,N64 

Insert a positional comma if the d value is omitted but the link-editing code is supplied.

Figure 45. MODS Control Statement Format

If an application has more than one exit, specify the exit-name parameter for each
exit. Up to 16 exits can be specified. Use commas to separate multiple exit-name
parameters.

Exit-Name Parameter (Required)

The exit-name parameter identifies the exit and provides additional information.
Replace 'exit-name' with an E followed by the appropriate exit number. The 16
valid exit-names are listed below.

2–80 Syncsort MFX Programmer’s Guide

 MODS

Sort Sort Sort or Merge

Copy
Phase 1 Phase 2 Phase 3

E11
E14
E15 E15
E16
E17
E18

Exit E21
E25
E27

Name E31 E31

E32 (merge only)
E35 E35
E37 E37
E38 E38
E39 E39
E61 E61

Table 17. Phases and Permissible Exits

The exit-name parameter also provides the following information about the exit.
r The r value specifies the name of the user exit routine. Any valid name
other than SYNCE15 or SYNCE35 is acceptable. (SYNCE15 and
SYNCE35 are reserved names.) If the exit routine resides in a library,
specify the member name or alias name for the r value. For an exit coded
in REXX, r represents the REXX exec name.
b The b value specifies the exact or estimated decimal number of bytes the
exit routine requires in main storage. This number should include any
additional main storage required by the exit (e.g., buffers, GETMAINs,
etc.). Specify an estimate (without an E before the value) if the exact
number is not known. This number should only include storage require-
ments below the 16-megabyte line.
REXX exits have some additional storage requirements. REXX system
modules and control blocks need 26K, and each EXEC that is called will
require 12K of storage. In addition to any variables that the EXEC uses,
all special MFX variables will require storage (including space for a
record).
d The d value identifies the DD statement name that specifies the library
in which the exit routine resides. The JCL must include a DD statement
specifying each library in which an exit routine resides. If the exit rou-
tine is to be placed in the input job stream, specify SYSIN for the d
value. (If more than one exit routine is included in SYSIN, the exit rou-
tines must be specified in ascending numerical order by exit name.)

Syncsort MFX Programmer’s Guide 2–81

MODS

For a Disk Sort, MAXSORT, or PARASORT, an exit routine that is a load

module residing in a library identified in a LINKLIB, STEPLIB or JOB-
LIB DD statement does not require a d value specification or a DD state-
ment defining a module library in the JCL. If the d value is omitted,
insert a positional comma to indicate the missing value. Do not specify
LINKLIB for the d value. LINKLIB is a reserved name.
The exit-name parameter also specifies link-editing codes: N, S, C, E, X, or T. If the
link-editing code is omitted, the installation setting determines whether or not the
exit will be link-edited. The delivered default is T; however, it may have been reset
to N at installation.
Ideally, exit routines should be designed so that they do not require link-editing
each time they are used. Link-editing consumes system resources and increases
sort/merge execution time.
When a link-editing code is specified, the name E10 is reserved and no Phase 1 exit
or E61 exit can use this name as a CSECT or ENTRY name. Similarly, the names
E20 and E30 are reserved and cannot be used by Phase 2 or Phase 3 exits.
N The N value specifies that link-editing is not required. Link-editing has
already taken place and MFX can directly invoke the routine.
S The S value specifies that link-editing is required. This value can only be
used for E11, E21 and E31 exits. The S value also indicates that the exit
routine can be link-edited separately from other exit routines specified
for the same phase.
C The C value identifies a COBOL exit routine. COBOL exits must be link-
edited before execution time. Only COBOL E15 and E35 exits can be
specified.
E The E value identifies a C exit routine. C exits must be link-edited before
execution time. Only C E15 and/or E35 exits can be specified.
X The X value identifies a REXX exit routine. Only REXX E15 and E35
exits can be specified.
T The T value specifies that MFX will dynamically link-edit the exit rou-
tine along with other routines specified for the same sort/merge phase.
N64 The N64 value specifies that the exit should be called using a 64-bit
parameter list. Register 13 will point to a Format 4 save area (See “Reg-
ister Conventions” on p. 7.4). This specification may only be used when
MFX is invoked via a 64-bit parameter list, and is only applicable to E15
and E35 exits. This does not affect the addressing mode in which the exit
will be called. The exit will be entered in the addressing mode indicated
by the linkage editor module attributes.

2–82 Syncsort MFX Programmer’s Guide

 MODS

Sample MODS Control Statement

MODS E15=(ADDREC1,600,MODLIB,N),E25=(ALTREC,500,SYSIN),
E35=(ADDREC2,600,MODLIB,C)

Figure 46. Sample MODS Control Statement

This sample MODS control statement specifies the following information:

• An E15 exit is the first exit routine. ADDREC1 is the member name of the
routine, which requires 600 bytes in main storage and resides in a library
referenced by the DD statement named MODLIB. The routine does not require
link-editing.
• An E25 exit is the second exit routine. ALTREC is the member name of the
routine which requires 500 bytes in main storage. The exit is included in the
SYSIN input stream. Because N is not specified, this routine will be link-edited.
• An E35 exit is the third exit routine. ADDREC2 is the member name of the
routine, which requires 600 bytes in main storage and resides in a library
referenced by the DD statement named MODLIB. This routine is a COBOL exit
which has been link-edited before execution time.
Examples of JCL-initiated applications with exit routines are illustrated in
“Chapter 4, JCL and Sample JCL/Control Statement Streams”.

Syncsort MFX Programmer’s Guide 2–83

OMIT

OMIT Control Statement

See “INCLUDE/OMIT Control Statement” on page 2-29 for an explanation of the
OMIT control statement.

2–84 Syncsort MFX Programmer’s Guide

 OUTFIL

OUTFIL Control Statement

The OUTFIL control statement describes the output file(s) and the processing to be
done on the output records. Use the OUTFIL statement to accomplish the following
tasks:
• Create multiple output files
• Create reports using the SortWriter facility
• Create and optionally e-mail an output file in a PDF, HTML or RTF format
• Reformat records after E35 processing

The Multiple Output Capability

Use the OUTFIL control statement to create multiple output files without making
multiple passes through the input data. The output files can be treated the same or
differently:
• The output files can contain the same or different records.
• The records in the output files can be identically or differently formatted.
• Whether the input files are fixed-length or variable-length, the output files may
be either.
Note that all the output files will be sequenced in the same way, as specified on the
SORT or MERGE control statement. If you need to sort the output files differently,
you should use Syncsort PipeSort, a Precisely product that works with Syncsort
MFX to reduce total elapsed time by generating multiple, differently sequenced
output files from a single read of the input data.
The OUTFIL parameters associated with this task are CONVERT, ENDREC,
FILES, FINDREP, FNAMES, FTOV, IFTHEN, INCLUDE/OMIT, NULLOFL,
OUTREC, OVERLAY, REPEAT, SAMPLE, SAVE, SPLIT, SPLITBY, SPLIT1R,
STARTREC, VLFILL, VLTRAIL, and VLTRIM.

The SortWriter Capability

The SortWriter capability of OUTFIL can produce completely formatted reports.
The report writing features, which can be specified differently for each output file,
can accomplish these tasks:
• Arrange the report into pages.
• Divide the report into sections.
• Format headers and trailers for sections, pages, and the complete report.
• Create multiple lines of output from each input record.
• Convert and edit numeric data.

Syncsort MFX Programmer’s Guide 2–85

OUTFIL

• Provide TOTAL and SUBTOTAL capabilities for data fields in a specific part of
a report.
• Provide MIN, MAX, AVG, SUBMIN, SUBMAX, and SUBAVG capabilities for
data fields in a specific part of a report.
• Provide COUNT and SUBCOUNT capabilities for records in a specific part of a
report.
Once formatted, output files can be assigned to any tape, disk, or unit record device
for subsequent printing.
The OUTFIL parameters associated with this task are BLKCCH1, BLKCCH2,
BLKCCT1, HEADER1, HEADER2, LINES, NODETAIL, REMOVECC,
SECTIONS, TRAILER1, and TRAILER2.

The PDF, HTML, RTF and E-mail Capability

An output file can be created in a PDF, HTML or RTF format. Any of these files can
be e-mailed as an attachment or attachments to one or more recipients. The
OUTFIL parameter associated with this task is OUTPUT.

The POST-E35 Reformatting Capability

Use the OUTFIL control statement if you need to reformat records after E35
processing. The OUTFIL parameters associated with this task are FINDREP,
IFOUTLEN, IFTHEN, OUTREC, OVERLAY, and PARSE.

OUTFIL Control Statement Format

The format for the OUTFIL control statement is illustrated on the following page.

2–86 Syncsort MFX Programmer’s Guide

 OUTFIL

OUTFIL [ACCEPT=n][,BLKCCH1][,BLKCCH2][,BLKCCT1]

 ,CONVERT  [ ,ENDREC=n ]
 
 ,VTOF 

 
,FILES =  fileid  ,FNAMES =  ddname 
( fileid 1 [ ,fileid 2 ]... )   ( ddname [ ,ddname ]... ) 
  1 2

[ ,FTOV ] [ ,HEADER1= ( field 1 [ ,field 2 ]... ) ] [ ,HEADER2= ( field 1 [ ,field 2 ]... ) ]

 ALL 
 NONE 
 
 ,INCLUDE    ,AND,  
 ,OMIT =  ,&,   [ ,IFTRAIL=(subparms) ]
   (c 1   c 2 ... ) 
  ,OR,  
  ,|,  
 

n   RC0 
   
,LINES =  ANSI  [ ,NODETAIL ] ,NOTMTOFL =  RC4 
 (ANSI,n)   RC16 
   

 RC0 
 
,NULLOFL =  RC4  [ , OUTPUT = ( subparm )]
 RC16 
 

 
   
  ,OUTREC  
 [ , PARSE=(subparm) ]  ,BUILD = ( field [ ,field ]... ) 
   1 2
 [ ,REMOVECC ]
,OVERLAY
   
 , IFTHEN= ( subparm ) [ ,IFTHEN = ( subparm )... ] [ ,IFOUTLEN =n ]
 , FINDREP=(subparm) 
 

n 
[ ,REPEAT=n ] ,SAMPLE=   [ ,SAVE ] [ ,SECTIONS= ( field 1 [ ,field 2 ]... ) ]
 (n,m) 

,SPLIT 
 ,SPLITBY=n  [ ,STARTREC=n ]
 ,SPLIT1R=n 

[ ,TRAILER1= ( field 1 [ ,field 2 ]... ) ] [ ,TRAILER2= ( field 1 [ ,field 2 ]... ) ]

[ ,VLFILL=f ] [ ,VLTRAIL=string ] [ ,VLTRIM=b ]

Figure 47. OUTFIL Control Statement Format

Syncsort MFX Programmer’s Guide 2–87

OUTFIL

FILES Parameter (Optional)

The FILES parameter connects the OUTFIL control statement with one or more
output files. The files specified on this parameter, along with any specified on the
FNAMES parameter, will constitute the ddnames to receive output for this
OUTFIL specification.
The format of the FILES parameter is illustrated in the following figure.

 fileid 
FILES =  
 (fileid 1 [,fileid 2 ] ...) 
where:
 OUT 
 
fileid =  x 
 xx 
 

Figure 48. FILES Parameter Format

The fileid identifies the output file and connects the OUTFIL control statement
with the corresponding SORTOUT, SORTOFx, or SORTOFxx DD statement. For
example, FILES=OUT connects the OUTFIL control statement with the
SORTOUT DD statement. Similarly, FILES=1 connects the OUTFIL control
statement with the SORTOF1 DD statement, and FILES=01 connects the OUTFIL
control statement with the SORTOF01 DD statement. The x can be any
alphanumeric character or special character allowed by JCL DD statements.
If multiple output files have identical specifications (that is, identical record
selection, record reformatting, and report writing specifications), the FILES and/or
FNAMES parameter can connect the OUTFIL control statement with more than
one DD statement. For example, FILES=(OUT,02,03) connects the OUTFIL control
statement with the SORTOUT, SORTOF02, and SORTOF03 DD statements. Such
a set of output files is termed an OUTFIL group.
If multiple output files have different specifications, then each file is specified on a
separate OUTFIL control statement with one FILES and/or FNAMES parameter
on each control statement.
If a SORTOUT ddname is defined in the JCL and does not appear in any FILES or
FNAMES specification, it will be written to without any OUTFIL processing. If an
inline E35 exit has been specified, OUTFIL is ignored.
If neither a FILES nor FNAMES parameter is specified on an OUTFIL control
statement, the default ddname of SORTOUT will be used. If a 4-byte ddname
prefix is in effect, the default SORTOUT ddname will be ppppOUT, where pppp is
the prefix; adding FILES=xx would connect to the ppppOFxx DD statement.

2–88 Syncsort MFX Programmer’s Guide

 OUTFIL

FNAMES Parameter (Optional)

The FNAMES parameter connects the OUTFIL control statement with one or more
output files. The files specified on this parameter, along with any specified on the
FILES parameter, will constitute the ddnames to receive output for this OUTFIL
specification.
The format of the FNAMES parameter is illustrated in the following figure.

 ddname 
FNAMES=  
 (ddname 1 [,ddname 2 ]...) 
Figure 49. FNAMES Parameter Format

ddname is a 1 to 8-character ddname that corresponds to a DD statement provided

in the JCL.
If multiple output files have identical specifications (that is, identical record
selection, record reformatting, and report writing specifications), the FNAMES
and/or FILES parameter can connect the OUTFIL control statement with more
than one DD statement. For example,
FNAMES=(FILE1OUT,FILE2OUT,FILE3OUT) connects the OUTFIL control
statement with the three listed DD statements. Such a set of output files is termed
an OUTFIL group.
If multiple output files have different specifications, then each file is specified on a
separate OUTFIL control statement with one FNAMES and/or FILES parameter
on each control statement.
If a SORTOUT ddname is defined in the JCL and does not appear in any FILES or
FNAMES specification, it will be written to without any OUTFIL processing. If an
inline E35 exit has been specified, OUTFIL is ignored.
If neither a FILES nor FNAMES parameter is specified on an OUTFIL control
statement, the default ddname of SORTOUT will be used. If a 4-byte ddname
prefix is in effect, the default SORTOUT ddname will be ppppOUT, where pppp is
the prefix.

INCLUDE/OMIT Parameter (Optional)

See “COND Parameter (Required)” on page 2-31 for a complete description of
comparisons and logical expressions.
Specify the INCLUDE or OMIT parameter to indicate which records are to be
included in or omitted from each output file. These parameters let you create
multiple output files which contain different records. The default is to include all
sorted or merged records in the output file.
The comparison determines which records are included or omitted. When no data
records are to be included in the output file(s) (when running a test, for example),
specify either INCLUDE=NONE or OMIT=ALL.

Syncsort MFX Programmer’s Guide 2–89

OUTFIL

Note: The location within the data records of the fields specified in the
INCLUDE/OMIT parameter will be based on the formatting of the record after pro-
cessing by an E15/E32 exit, the INREC control statement, the OUTREC control
statement, and an E35 exit, but before processing due to the OUTREC and/or
report writing parameters of the OUTFIL control statement.
The following four parameters (STARTREC, ENDREC, SAVE, and SPLIT) are
related to the previous parameter (INCLUDE/OMIT) in that they specify records to
be included for OUTFIL processing. However, these four options specify records in
bulk rather than through a comparison condition.

REPEAT Parameter (Optional)

The REPEAT=n parameter enables each output record to be written multiple
times. n specifies the number of times each OUTFIL output record is written. The
minimum value for n is 2.
REPEAT can be used with the OUTFIL OUTREC multiline feature (designated by
a / in the OUTREC specification). When this is done, each line will be written n
times defined by the OUTREC specification. All occurrences of the first line will be
written followed by all the occurrences of the second, and so on.
The REPEAT parameter cannot be used with IFTRAIL, LINES, HEADER1,
TRAILER1, HEADER2, TRAILER2, SECTIONS, and NODETAIL.

ACCEPT Parameter (Optional)

The ACCEPT=n parameter is used to limit the number of records processed for the
OUTFIL group. Records entering OUTFIL processing that are not excluded by any
of the INCLUDE, OMIT, SAMPLE, STARTREC, or ENDREC parameters are
included in the ACCEPT count, and no further records are processed after n
records have been included. If ENDREC has also been specified, processing will
stop when either one has been satisfied.

STARTREC Parameter (Optional)

Use the STARTREC=n parameter to specify the record number n of the first record
to be processed by the OUTFIL specification in effect. All records prior to the
specified record will be ignored for the OUTFIL group. The record number is
determined by the sequence of records presented for OUTFIL processing.
For more information, see “SAMPLE Parameter (Optional)” on page 2-91.

ENDREC Parameter (Optional)

Use the ENDREC=n parameter to specify the record number n of the last record to
be processed by the OUTFIL specification in effect. All records after the specified
record will be ignored for the OUTFIL group. The record number is determined by
the sequence of records presented for OUTFIL processing.

2–90 Syncsort MFX Programmer’s Guide

 OUTFIL

If ACCEPT has also been specified, processing will stop when either one has been
satisfied.

SAMPLE Parameter (Optional)

The SAMPLE=n and SAMPLE=(n,m) parameters allow the selection of a sample of
records from an OUTFIL group. A specific interval and number of records in that
interval can be specified. The sample process will take place within the range of
records specified by STARTREC or ENDREC if they are specified. SAMPLE=n and
SAMPLE=(n,m) are mutually exclusive.
The sample consists of the first m records in every nth interval. n specifies the
interval size. The minimum value for n is 2 (sample every other record).
m specifies the number of records to be processed in each interval. The minimum
value for m is 1 (process the first record in each interval). If m is not specified, 1 is
used for m. If m is specified, it must be less than n.

SAVE Parameter (Optional)

Use SAVE to include records for OUTFIL processing that have not been included in
any other OUTFIL group.
If SAVE is specified on more than one OUTFIL group, then each of these OUTFIL
groups get the records that were discarded from all other OUTFIL groups that do
not have SAVE.
The OUTFIL INCLUDE/OMIT parameter is mutually exclusive with the SAVE
parameter. Only one of these parameters can be specified for an OUTFIL group.
Note that if the SORTOUT data set has not been associated with any OUTFIL
control statement but is present in the JCL, the SORTOUT data set will receive a
copy of all records prior to OUTFIL processing. This does not affect the SAVE
operation, since SAVE is only pertinent to other OUTFIL group specifications.

SPLIT Parameter (Optional)

The SPLIT parameter of the OUTFIL control statement causes output records to be
distributed in rotation among files in an OUTFIL group.
In the normal case, when the SPLIT parameter is not used, the output files in the
group will contain the same records. SPLIT distributes the output records. The
following OUTFIL control statement will distribute records among three output
files:

OUTFIL FILES=(01,02,03),SPLIT

Figure 50. Sample OUTFIL Control Statement with SPLIT

Syncsort MFX Programmer’s Guide 2–91

OUTFIL

For the above example, the first record will be written to the SORTOF01 data set;
the second, to SORTOF02; the third, to SORTOF03. The fourth record will be
written to SORTOF01 again, and so on in round-robin fashion.
The OUTFIL control statement can contain an INCLUDE/OMIT and an OUTREC
parameter, in which case the selected and reformatted subset of records will be
distributed among the output files.
SPLIT, SPLITBY=n and SPLIT1R=n are mutually exclusive. SPLITBY=1 is
equivalent to SPLIT.
Note that the SPLIT parameter cannot be used with any report writing
(SortWriter) functions. Specifically, report writing parameters (HEADERn,
TRAILERn, SECTIONS, LINES, NODETAIL, IFTRAIL) cannot be specified on the
OUTFIL control statement that defines the output group.
SPLIT can be used with BatchPipes/MVS; that is, the output records can be
distributed among BatchPipes/MVS data sets.

SPLITBY Parameter (Optional)

The SPLITBY=n parameter writes groups of records in rotation among multiple
output data sets and distributes multiple records at a time among the OUTFIL
data sets. n specifies the number of records to split by. The minimum value for n is
1.
The SPLITBY parameter is similar to SPLIT, but SPLITBY can be used to rotate
by a specified number of records rather than by one record, for example, records 1-
10 to the first OUTFIL data set, records 11-20 to the second OUTFIL data set, and
so on.
For example, if SPLITBY=10 is specified for an OUTFIL group with three data
sets:
• The first OUTFIL data set in the group receives records 1-10, 31-40, and so on.
• The second OUTFIL data set in the group receives records 11-20, 41-50, and so
on.
• The third OUTFIL data set in the group receives records 21-30, 51-60, and so
on.
SPLIT, SPLITBY=n and SPLIT1R=n are mutually exclusive. SPLITBY=1 is
equivalent to SPLIT.
Note that the SPLITBY parameter cannot be used with any report writing
(SortWriter) functions. Specifically, report writing parameters (HEADERn,
TRAILERn, SECTIONS, LINES, NODETAIL, IFTRAIL) cannot be specified on the
OUTFIL control statement that defines the output group.

2–92 Syncsort MFX Programmer’s Guide

 OUTFIL

SPLIT1R Parameter (Optional)

The SPLIT1R=n parameter of the OUTFIL control statement causes all output
records to be grouped and distributed among files in an OUTFIL group in a single
rotation to maintain contiguity. The first n records are written to the first output
file followed by the next n records written to the next output file and so on, with the
remaining records written to the last output file regardless of n. The value of n
must be specified.
SPLIT, SPLITBY=n and SPLIT1R=n are mutually exclusive.
Note that the SPLIT1R=n parameter cannot be used with any report writing
(SortWriter) functions. Specifically, report writing parameters (HEADERn,
TRAILERn, SECTIONS, LINES, NODETAIL, IFTRAIL) cannot be specified on the
OUTFIL control statement that defines the output group.

OUTFIL FILES=(01,02,03),SPLIT1R=10

Figure 51. Sample OUTFIL Control Statement with SPLIT1R=n

For the above example, given 36 records to be distributed in a single rotation,

records 1 to 10 will be written to the first data set; records 11 to 20 will be written
to the second data set; and records 21 to 36 will be written to the third data set.

OUTREC/BUILD Parameter (Optional)

The OUTREC parameter indicates how the records are to be formatted in each
output file. (BUILD is an alias for OUTREC.) This parameter lets you create
multiple output files which contain differently formatted records.
When the records in all multiple output files are formatted and edited identically, it
is more efficient to specify a single OUTREC control statement rather than several
OUTREC parameters.
The OUTREC parameter reformats the records that are to be included in the
output file(s) after E35 processing, if specified. If no additional reformatting is
required, omit this parameter.
All references to field positions specified in the OUTREC parameter refer to the
record after processing by an E15 exit, the INREC control statement, the OUTREC
control statement, and an E35 exit but before insertion of ANSI control characters.
The format of the OUTREC parameter is illustrated below.

OUTREC=(field1[,field2]...)

Figure 52. OUTREC Parameter Format

The format of the OUTFIL OUTREC parameter is generally identical to the format
of the FIELDS parameter of the OUTREC control statement. (See the subsections

Syncsort MFX Programmer’s Guide 2–93

OUTFIL

dealing with the FIELDS parameter in “OUTREC Control Statement” on page 2-

134.) Note, however, that FIELDS= is not used with OUTFIL OUTREC. In
addition, OUTFIL OUTREC accepts the / subparameter and can be used with the
VLFILL parameter.
[n]/ The / subparameter indicates the end of a line and can be used to cre-
ate multiple output lines from a single input record. Multiple slashes
(coded
//.../ or n/) can be used to specify leading, trailing, or embedded blank
lines. At the beginning or end of the OUTREC parameter, n/ pro-
duces n blank lines. Embedded within the OUTREC parameter, n/
produces n-1 blank lines.
The / subparameter is most useful for its ability to accommodate
records whose lengths exceed the width of the physical page. For an
example of the / subparameter, see “Printing Input Records on Multi-
ple Output Lines” on page 3-39.
The / subparameter may not be used when LINES=ANSI or
LINES=(ANSI,n) has also been specified on the OUTFIL control
statement.

IFTHEN Parameter (Optional)

The IFTHEN parameter employs conditional logic, which enables you to reformat
your records based on specified criteria. Multiple IFTHEN parameters may be
specified within the same control statement and are processed sequentially. The
IFTHEN parameter may be used within the INREC, OUTREC, and OUTFIL
control statements. See “IFTHEN Parameter (Optional)” on page 2-198 for a
complete description.

PARSE Parameter (Optional)

The PARSE parameter is used to extract variable-position and variable-length
fields from records and place the resultant data into fixed-length parsed fields. See
“PARSE Parameter (Optional)” on page 2-209.

IFOUTLEN Parameter (Optional)

The IFOUTLEN parameter overrides the maximum record length, which is
automatically set by the IFTHEN parameter, and changes it to a specified value.
The IFOUTLEN parameter may only be used in conjunction with the IFTHEN
parameter. See “IFOUTLEN Parameter (Optional)” on page 2-208 for a complete
description.

FINDREP Parameter (Optional)

The FINDREP parameter provides the ability to find and replace one or more
constants in a record. A constant to be searched for can be specified as a character

2–94 Syncsort MFX Programmer’s Guide

 OUTFIL

or hexadecimal string and its replacement constant can be either a character,

hexadecimal or null string.
See “FINDREP Parameter (Optional)” on page 2-193 for details on its use.

OVERLAY Parameter (Optional)

The OVERLAY parameter enables you to change particular columns and add fields
to the end of a record without rebuilding the entire record. When using the
OVERLAY parameter you only need to specify the columns you want to change.
The rest of the input record remains unchanged. See “OVERLAY Parameter
(Optional)” on page 2-208 for a complete description.

VLFILL Parameter (Optional)

The VLFILL parameter is used in conjunction with OUTREC or OUTREC
CONVERT to specify a fill byte to be used for any missing p,l field bytes.
The VLFILL parameter has two functions:
• It enables a variable-length OUTFIL OUTREC non-CONVERT application to
continue processing when there is an input record with missing field bytes in a
p,l field specification.
• It provides a means to override the default fill byte used in an OUTFIL
OUTREC CONVERT application when there are missing bytes in a p,l field
specification.
In the first instance, if VLFILL has not been specified the application will
terminate with the critical error WER244A. In the second case, by default, spaces
will be used for missing field bytes.
f specifies a byte to be used for missing field bytes. f can be specified as either a
character or hexadecimal value. Specify either C'x' where x is a single EBCDIC
character or X'hh' where hh represents a hexadecimal digit pair (00-FF).
Note: VLFILL is ignored when the FTOV parameter is used or when the OUTREC
parameter is not used. VLFILL may not be used with the IFTRAIL parameter.

CONVERT Parameter (Optional)

The CONVERT parameter is used in conjunction with the OUTREC parameter to
convert variable-length records to fixed-length records.
The records do not require an RDW and will be written to the output file(s) with a
RECFM of F or FB. When using CONVERT, you no longer need to apply the rules
for “Specifying the FIELDS parameter for Variable-Length Records” found in the
description of the OUTREC control statement.
You cannot specify the variable portion of the input records (position without
length) when using CONVERT. All other p,l data fields that are not present will be

Syncsort MFX Programmer’s Guide 2–95

OUTFIL

filled with blanks by default. The OUTFIL VLFILL parameter can be used to
specify a different fill byte for any missing fields (see above description).
An E35 exit may process converted fixed-length records by using CONVERT on the
OUTREC statement rather than on the OUTFIL statement.
Note: If CONVERT is specified, the OUTREC or BUILD parameter must also be
specified. CONVERT cannot be used with the IFTRAIL, FTOV or VTOF parameter.

VTOF Parameter (Optional)

VTOF is equivalent to CONVERT. See “CONVERT Parameter (Optional)” on
page 2-95.

FTOV Parameter (Optional)

The FTOV parameter converts fixed-length input records to variable-length output
records.
FTOV can be used both with and without the OUTREC parameter. When FTOV is
used with the OUTREC parameter, the variable-length record is created from the
specified fields of the fixed-length record. When FTOV is not used with the
OUTREC parameter, the variable-length record is created from the whole fixed-
length record.
Note: FTOV cannot be used with IFTRAIL, CONVERT or VTOF. If the input
record is variable-length, FTOV, if specified, will be ignored. FTOV can be used
with the VLTRIM parameter to delete pad bytes at the end of a record, or with the
VLTRAIL parameter to add bytes at the end of a record.
For an example of an OUTFIL control statement that uses the FTOV parameter,
see Figure 71 on page 2-136.

VLTRAIL Parameter (Optional)

The VLTRAIL parameter allows variable-length output records to be appended
with a character string (C’text’) or a hexadecimal string (X’hh..’).

 C'text' 
VLTRAIL =  
 X'hh..' 

Figure 53. VLTRAIL Format

The specified string has a maximum length of 50 bytes and will be added at the end
of both variable-length data records and header/trailer records. VLTRAIL is
ignored for fixed-length output records.

2–96 Syncsort MFX Programmer’s Guide

 OUTFIL

The LRECL of the output data set must be large enough to accommodate the
increase in record length, or MFX will terminate with a WER167A error message.
This may require you to specify an LRECL on the output DD statement(s).
VLTRAIL processing is performed after processing for the optional VLTRIM
parameter. VLTRAIL is mutually exclusive with IFTRAIL.

VLTRIM Parameter (Optional)

The VLTRIM parameter defines a byte to be deleted from the end of a variable-
length record. All prior occurrences of this byte will also be deleted until a byte that
is not equal to the trim byte is found. The resulting records are decreased in record
length. However, VLTRIM will not delete the first data byte, the ANSI carriage
control character, or the Record Descriptor Word (RDW).
The format of the VLTRIM parameter is illustrated below.

VLTRIM=b

Figure 54. VLTRIM Format

b specifies the byte to be deleted from the end of the record. b can be specified as
either a character or hexadecimal value. Specify either C'x' where x is a single
EBCDIC character or X'hh' where hh represents a hexadecimal digit pair (00-FF).
Note: VLTRIM is ignored if used with fixed-length output records. VLTRIM may
not be used with the IFTRAIL parameter.
For an example of an OUTFIL control statement that uses the VLTRIM parameter,
see
Figure 71 on page 2-136.

BLKCCH1 Parameter (Optional)

The BLKCCH1 parameter of the OUTFIL control statement prevents a page eject
at the start of HEADER1, the report header. In the first line of HEADER1, a blank
character replaces the ANSI carriage control character '1'. When specifying
BLKCCH1, HEADER1 also must be specified; otherwise, it is ignored.

BLKCCH2 Parameter (Optional)

The BLKCCH2 parameter of the OUTFIL control statement prevents a page eject
at the start of the first HEADER2, the page header. In the first line of HEADER2, a
blank character replaces the ANSI carriage control character '1'. When specifying
BLKCCH2, HEADER2 also must be specified; otherwise, it is ignored.

Syncsort MFX Programmer’s Guide 2–97

OUTFIL

BLKCCT1 Parameter (Optional)

The BLKCCT1 parameter of the OUTFIL control statement prevents a page eject
at the start of TRAILER1, the report trailer. In the first line of TRAILER1, a blank
character replaces the ANSI carriage control character '1'. When specifying
BLKCCT1, TRAILER1 also must be specified; otherwise, it is ignored.

HEADER1/HEADER2 Parameters (Optional)

The SortWriter facility provides three types of headers:
• HEADER1, the report header
• HEADER2, the page header
• HEADER3, the section header.
HEADER1 and HEADER2 are parameters of the OUTFIL control statement.
HEADER3 is a subparameter of OUTFIL’s SECTIONS parameter. Refer to
“SECTIONS Parameter (Optional)” on page 2-118 for an explanation of how to
specify HEADER3.
The three types of headers function independently of each other. Each serves a
different purpose.
• HEADER1 provides a header or a possible title page for the entire report. It
appears only once at the beginning of the report on its own page.
• HEADER2 provides a page header or a running head for each page defined by
the LINES parameter. It appears at the beginning or top of each page.
• HEADER3 provides a section header that appears at the beginning of each
specified section and, optionally, at the top of each page (or directly below any
HEADER2).
The chart below illustrates the format for HEADERs. The field entries represent
the subparameters that can be specified for each HEADER entry.

HEADER1=(field1[,field2]...)
HEADER2=(field1[,field2]...)
HEADER3=(field1[,field2]...)

Figure 55. HEADER Parameter Format

The following HEADER Subparameters Format chart illustrates and defines the
available subparameters. Each subparameter constitutes a separate field of the
HEADER.

2–98 Syncsort MFX Programmer’s Guide

 OUTFIL

 [n] X 
 [n] X'hhhh...hh' 
 [n] 'literal string' 
 
 [n] / 
 p,l 
 &DATE [ { ± }nnnn ] 
 
 &DATE=(m m
1 2 3 4m m ) [ { ± }nnnn ] 
 &DATENS=(xyz) [ { ± }nnnn ] 
 
 &YDDD=(m m
1 2 3 m ) [ { ± }nnnn ] 
 &YDDDNS=(m 1 m 2 ) [ { ± }nnnn ] 
 
[ c: ]  &TIME 
 &TIME=(hp) 
 &TIMENS=(tt) 
 
 &PAGE 
   
 f o [,LENGTH=n] 
   
 TO = f o [,LENGTH=n] 
   
 &PAGE=(   Mm   ) 
    [,SIGNS=(...)] [,LENGTH=n]  
 
EDIT=(...) 
 
   
  M0  


Figure 56. HEADER Subparameters Format

c: Use the c: subparameter to define the column in which the

specified field should begin. The c: value will ignore the car-
riage control character and, for variable-length records, the
RDW. EBCDIC blanks will precede the specified column
when needed.
n Used in conjunction with the X, X'hh..hh', 'literal string',
and / subparameters, the n value defines the number (1-
4095) of occurrences for each entry.
X Use the X subparameter to define the number of spaces. It
must be coded to the immediate right of the n value, if spec-
ified. For more than 4095 spaces, two or more nX values
should be specified.
X'hhhh...hh' Use the X'hhhh...hh' entry to specify that a hexadecimal
string should be inserted in the header. (Each hh pair is 1
byte of output.) Specify the number of occurrences by coding
n immediately before it.
'literal string' Use the 'literal string' subparameter to define a literal
string. Specify the number of occurrences by coding n imme-
diately before it. An apostrophe within a literal string must
be specified as a double apostrophe; for example, –
'O"Leary'.

Syncsort MFX Programmer’s Guide 2–99

OUTFIL

/ Use the / subparameter to indicate the end of a line, force a

carriage return, and separate text lines of a header. Multi-
ple slashes (//.../ or n/) can be used to specify leading, trail-
ing, or embedded blank lines. At the beginning or end of a
header, n/ produces n blank lines. Within a header, n/ pro-
duces n-1 blank lines.
p,l Use the p and l subparameters to include a field (or fields)
within a record in the header. For a HEADER1, the field(s)
will be extracted from the first record in a file; for a HEAD-
ER2, the field(s) will be extracted from the first record on a
page; for a HEADER3, the field(s) will be extracted from the
first record in a section. p is the starting position of the field
in the record; l is the length in bytes (1-32752) of the field.
Any number of fields can be specified. (Contiguous fields
within a record can be specified with a single p,l entry, but
their combined length cannot exceed 32752 bytes.) The spec-
ified field(s) should be a character or alphanumeric string or
a number in printable format, and the field(s) cannot be
converted or edited.
&DATE [{±}nnnn] The &DATE subparameter specifies the current system
date or date with offset and requires 8 bytes to display
mm/dd/yy.
Optionally, you can create an offset of the current date. The
offset takes the form {±}nnnn, where '+' indicates a date
after the current date and '–' indicates a date before the cur-
rent date. 'nnnn' is the date offset. The range is 0 to 9999,
which represents the number of days to be added or sub-
tracted from the current date.
&DATE=(m1m2m3m4)[{±}nnnn] This form of the &DATE subparameter gener-
ates the current system date or date with offset and controls
the formatting of the date. You can specify the position of
the year, month, and day, specify a separator character, and
choose between 2-digit and 4-digit year representation.
The positions m1 through m4 represent masks used to for-
mat the date. To specify the position of the month, day, and
year, replace the m1, m2, and m3 positions, in any order, with
M for the month (01-12), D for the day (01-31), and either Y
or 4 for the year (where Y is a 2-digit year and 4 is a 4-digit
year). Replace the m4 position with a separator character.
For example, to print the date with the form yy-mm-dd,
specify &DATE=(YMD-). For December 31, 1999, the date
would appear as “99-12-31.”

2–100 Syncsort MFX Programmer’s Guide

 OUTFIL

The field for this form of &DATE requires 8 bytes for a 2-

digit year representation and 10 bytes for a 4-digit year. The
M, D, and Y or 4 may only appear once in the mask. All four
positions must be specified.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-100 for a description.
&DATENS=(xyz)[{±}nnnn] This form of the &DATENS subparameter specifies
that the current date or date with offset is to appear in the
report record in the form 'xyz', where x, y, and z indicate the
order in which the month, day, and year are to appear and
whether the year is to appear as two or four digits. For x, y,
and z, use M to represent the month (01-12), D to represent
the day (01-31), Y to represent the last two digits of the year
(for example, 02), or 4 to represent the four digits of the year
(for example, 2002). M, D, and Y or 4 can each be specified
only once.
For example, &DATENS=(DMY) would produce a date of
the form 'ddmmyy' which on March 29, 2002, would appear
as '290302'. &DATENS=(4MD) would produce a date of the
form 'yyyymmdd' which on March 29, 2002, would appear as
'20020329'. x, y, and z must be specified.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-100 for a description.
&YDDD=(m1m2m3)[{±}nnnn] This form of the &YDDD subparameter specifies
that the current date or date with offset is to appear in the
report record in the form of a year and day. You can specify
the position of the year and day, specify a separator charac-
ter, and choose between 2-digit and 4-digit year representa-
tion. The positions m1 through m3 represent masks used to
format the date. To specify the position of the year and day,
replace the m1 and m2 positions (in either position) with D
for day (001-366) and either Y or 4 for the year (where Y is a
2-digit year and 4 is a 4-digit year). Replace the m3 position
with a separator character.
For example, to print the date in the form yyyy/ddd, specify
&YDDD=(4D/). For March 29, 2005, the date would appear
as 2005/088.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-100 for a description.
&YDDDNS=(m1m2)[{±}nnnn] This form of the &YDDDNS subparameter speci-
fies that the current date or date with offset is to appear in
the report record in the form of a year and day. You can
specify the position of the year and day and choose between

Syncsort MFX Programmer’s Guide 2–101

OUTFIL

2-digit and 4-digit year representation. The positions m1

and m2 represent masks used to format the date. To specify
the position of the year and day, replace the m1 and m2 posi-
tions (in either position) with D for day (001-366) and either
Y or 4 for the year (where Y is a 2-digit year and 4 is a 4-
digit year).
For example, to print the date in the form dddyy, specify
&YDDDNS=(DY). For March 29, 2005, the date would
appear as 08805.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-100 for a description.
&TIME The &TIME subparameter specifies the current time of day
and requires 8 bytes to display hh:mm:ss, where hh is in 24-
hour format.
&TIME=(hp) This form of the &TIME subparameter generates the cur-
rent system time of day and controls the formatting of the
time. You can print the time in 24-hour or 12-hour format
and specify the separator character between the hours, min-
utes, and seconds.
The format for 24-hour time is hhpmmpss, where hh rep-
resents the hour (00-23), mm represents minutes (00-59), ss
represents seconds (00-59), and p represents the separator
character as specified by p in the &TIME=hp subparameter.
The format for 12-hour time is hhpmmpss nn, where hh rep-
resents the hour (01-12), mm represents minutes (00-59), ss
represents seconds (00-59), and p represents the separator
character as specified by p in the &TIME=hp subparameter.
The nn is “am” or “pm” as appropriate.
To select 12-hour mode specify h as 12; to select 24-hour
mode specify h as 24. The p specification represents the
character to use as a separator. For example, to display the
time in a 12-hour format with a period as a separator, spec-
ify &TIME=(12.). At 22:43:23 hours, the time would appear
as “10.43.23 pm.”
The field for this form of the &TIME subparameter requires
8 bytes for the 24-hour format and 11 bytes for the 12-hour
format.
&TIMENS=(tt) This form of the &TIMENS subparameter specifies that the
current time is to appear in the report record in the form
'hhmmss' (24-hour time) or 'hhmmss xx' (12-hour time). If tt
is 24, the time is to appear in the form 'hhmmss' (24-hour

2–102 Syncsort MFX Programmer’s Guide

 OUTFIL

time) where hh represents the hour (00-23), mm represents

the minutes (00-59), and ss represents the seconds (00-59).
For example, &TIMENS=(24) would produce a time of the
form 'hhmmss' which at 08:25:13 pm would appear as
'202513'. If tt is 12, the time is to appear in the form
'hhmmss xx' (12-hour time) where hh represents the hour
(01-12), mm represents the minutes (00-59), ss represents
the seconds (00-59), and xx is either 'am' or 'pm'. For a sec-
ond example, &TIMENS=(12) would produce a time of the
form 'hhmmss xx' which at 08:25:13 pm would appear as
'082513 pm'.
&PAGE The &PAGE subparameter sequentially numbers logical
pages of the output report and requires 6 bytes. It produces
a 6-digit sequential page number, right justified with lead-
ing zeros suppressed. &PAGE is ignored for HEADER1.
&PAGE=(…) This form is similar to the &PAGE subparameter except
that a 15-digit page number will be provided for display
with editing or conversion to another data format.
The following describes the &PAGE subparameters that
control format conversion or printable display with edit:
fo Use this subparameter to define the out-
put
TO=fo numeric data format of an expression.
When fo is specified, mask Mm, EDIT,
and SIGNS cannot be specified. Indicate
the desired format of the output field by
replacing fo with BI, CSF/FS, FD, FI,
PD, or ZD. See “How to Convert
Numeric Data” on page 2-157 for the
default lengths of these fields. See
“LENGTH=n Subparameter” on page 2-
178 for how this default may be
changed.
TO=fo is equivalent to fo and in general,
there is no reason to use the TO= form.
However, if you are using a data dictio-
nary symbol in your control statement,
you should use the TO=fo form to avoid
ambiguities with certain types of data
conversions. See the section “INREC,
OUTREC, OUTFIL TO Subparameter”
on page 13-23.

Syncsort MFX Programmer’s Guide 2–103

OUTFIL

Mm Use the Mm subparameter to indicate

that one of the 27 MFX-provided editing
masks, M0-M26, is to be used. Replace
'm' with the mask number. For details,
see “Mm Subparameter (Editing
Masks)” on page 2-179.
EDIT=(pattern) Use the EDIT subparameter to specify
that a user-provided editing mask
should be used to format the output
fields. For details, see“EDIT Subparam-
eter” on page 2-177 .
SIGNS=(s1,s2,s3,s4) Use the SIGNS subparameter to specify
the signs that will appear before or after
the edited number. For details, see
“SIGNS Subparameter” on page 2-181.
LENGTH=n Use the LENGTH subparameter to alter
the length of the output field. This is
normally determined by the number of
numeric digits and either the data for-
mat or the edit pattern and format of
the edited field. For details, see
“LENGTH=n Subparameter” on page 2-
178.

Rules for Specifying HEADER Subparameters

Observe the following guidelines when you specify HEADER subparameters:
• Separate subparameters with commas, except between c: and another
subparameter. Commas are optional for the / subparameter.
• Enclose literals in single quotes.
• Specify blank fields of n bytes as nX.
• For fixed-length records, headings specified with fewer blanks than the logical
record length (LRECL) of the output record are automatically padded on the
right with blanks.
• If a header length exceeds the logical record length (LRECL) of the output
record, MFX will issue the WER116A error message. If you do not wish to
shorten the header, you can lengthen the record. For fixed-length output, use
the OUTREC control statement or the OUTREC parameter to expand the
output record length so that it is at least as long as the longest header (and
trailer). For example, if the longest header is 115 characters and the output
record length is 80 bytes, use the OUTREC control statement or the OUTREC
parameter to insert a blank in position 115 of the output record. This will cause
bytes 81 through 115 to be padded with blanks. For variable-length records, it is

2–104 Syncsort MFX Programmer’s Guide

 OUTFIL

easiest to just change the LRECL on the output data set DD statement to
match the length of the longest header (and trailer).
• HEADERn may not be used with the IFTRAIL parameter.

TRAILER Parameters (Optional)

The SortWriter facility provides three types of trailers:
• TRAILER1, the report trailer
• TRAILER2, the page trailer
• TRAILER3, the section trailer.
TRAILER1 and TRAILER2 are parameters of the OUTFIL control statement;
TRAILER3 is a subparameter of OUTFIL’s SECTIONS parameter. Refer to
“SECTIONS Parameter (Optional)” on page 2-118 for an explanation of how to
specify TRAILER3.
The three types of trailers function independently of each other. Each serves a
different purpose:
• TRAILER1 provides a trailer or a possible summary for the entire report. It
appears only once at the end of the report on its own page.
• TRAILER2 provides a page trailer for each page defined by the LINES
parameter. It appears at the end of each page.
• TRAILER3 provides a section trailer that appears at the end of each specified
section and serves as a conclusion or summary for that section.
TRAILER1, TRAILER2, and TRAILER3 also provide TOTAL, SUBTOTAL, MIN,
SUBMIN, MAX, SUBMAX, AVG, SUBAVG, COUNT, SUBCOUNT, COUNT15, and
SUBCOUNT15 capabilities at report, page, and section levels.
The chart below illustrates the format for TRAILERs. Its field entries represent the
subparameters that can be specified for each TRAILER entry.

TRAILER1=(field1[,field2]...)
TRAILER2=(field1[,field2]...)
TRAILER3=(field1[,field2]...)

Figure 57. TRAILER Parameter Format

The following TRAILER Subparameters Format chart illustrates and defines the
available subparameters. Each subparameter constitutes a separate field of the
TRAILER.

Syncsort MFX Programmer’s Guide 2–105

OUTFIL

 [n] X 
 [n] X'hhhh...hh' 
 
 [n] 'literal string' 
 [n] / 
 p,l 
 
 &DATE [ { ± }nnnn ] 
 &DATE= ( m 1 m 2 m 3 m 4 ) [ { ± }nnnn ] 
 
 &DATENS=(xyz) [ { ± }nnnn ] 
 &YDDD=(m 1 m 2 m 3 ) [ { ± }nnnn ] 
 &YDDDNS=(m m ) [ { ± }nnnn ] 
 1 2 
 &TIME 
 &TIME=(hp) 
 
 &TIMENS=(tt) 
 &PAGE 
 
  f o [,LENGTH=n]  
   
  TO = f o [,LENGTH=n]  
   
 &PAGE=(  Mm   ) 
 
 EDIT=(...)  [,SIGNS=(...)] [,LENGTH=n]  
    
  M0  
   
 
  TOTAL/TOT  
   , f [,LENGTH=n]  
  SUBTOTAL/SUB   o
 
[c:]   MIN   ,TO = f o [,LENGTH=n]  
  SUBMIN    
  = (p,l,f  ,Mm    )
 MAX  
 ,EDIT=(...)  [,SIGNS=(...)] [,LENGTH=n]  
  SUBMAX      
  AVG   ,M0  
    
 SUBAVG  
 
 COUNT 
   
 f o [,LENGTH=n] 
   
 TO = f o [,LENGTH=n] 
     
 COUNT  + nnn =(   Mm   ) 
  -    [,SIGNS=(...)] [,LENGTH=n]  
   EDIT=(...)   
   
  M0  
 
 SUBCOUNT 
 
  f o [,LENGTH=n]  
   
  TO = f o [,LENGTH=n]  
   
 SUBCOUNT=(   Mm  ) 
   EDIT=(...)  [,SIGNS=(...)] [,LENGTH=n]  
    
  M0  
   
 
 COUNT15 
 SUBCOUNT15 

Figure 58. TRAILER Subparameters Format

2–106 Syncsort MFX Programmer’s Guide

 OUTFIL

c: Use the c: subparameter to define the column in which the

specified field should begin. The c: value will ignore the car-
riage control character and, for variable-length records, the
RDW. EBCDIC blanks will precede the specified column
when needed.
n Used in conjunction with the X, X'hh..hh', 'literal string',
and / subparameters, the n value defines the number (1-
4095) of occurrences for each entry.
X Use the X subparameter to define the number of spaces. It
must be coded to the immediate right of the n value, if spec-
ified. For more than 4095 spaces, two or more nX values
should be specified.
X'hhhh...hh' Use the X'hhhh...hh' entry to specify that a hexadecimal
string should be inserted in the header. (Each hh pair is 1
byte of output.) Specify the number of occurrences by coding
n immediately before it.
'literal string' Use the 'literal string' subparameter to define a literal
string. Specify the number of repetitions by specifying n
immediately before it. An apostrophe within a literal string
must be specified as a double apostrophe; for example, –
'O"Leary'.
/ Use the / subparameter to indicate the end of a line, force a
carriage return, and separate text lines of a trailer. Multiple
slashes (coded //.../ or n/) can be used to specify leading,
trailing, or embedded blank lines. At the beginning or end-
ing of a trailer, n/ produces n blank lines. Within a trailer, n/
produces n-1 blank lines.
p,l Use the p and l subparameters to include a field (or fields)
within a record in the trailer. For a TRAILER1, the field(s)
will be extracted from the last record in a file; for a TRAIL-
ER2, the field(s) will be extracted from the last record on a
page; for a TRAILER3, the field(s) will be extracted from the
last record in a section. p is the starting position of the field
in the record; l is the length in bytes (1-32752) of the field.
Any number of fields can be specified. (Contiguous fields
within a record may be specified with a single p,l entry, but
their combined length may not exceed 32752 bytes.) The
specified field(s) should be a character or alphanumeric
string, or a number in printable format, and the field cannot
be converted or edited.
If any variable-length record contains only a portion of the
bytes in a specified field, those bytes will be included in the
trailer and blanks will be substituted for the missing bytes.

Syncsort MFX Programmer’s Guide 2–107

OUTFIL

&DATE [{±}nnnn] The &DATE subparameter specifies the current system

date or date with offset and requires 8 bytes to display
mm/dd/yy.
Optionally, you can create an offset of the current date. The
offset takes the form {±}nnnn, where '+' indicates a date
after the current date and '–' indicates a date before the cur-
rent date. 'nnnn' is the date offset. The range is 0 to 9999,
which represents the number of days to be added or sub-
tracted from the current date.
&DATE=(m1m2m3m4)[{±}nnnn] This form of the &DATE subparameter gener-
ates the current system date or date with offset and controls
the formatting of the date. You can specify the position of
the year, month, and day, specify a separator character, and
choose between 2-digit and 4-digit year representation.
The positions m1 through m4 represent masks used to for-
mat the date. To specify the position of the month, day, and
year, replace the m1, m2, and m3 positions, in any order, with
M for the month (01-12), D for the day (01-31), and either Y
or 4 for the year (where Y is a 2-digit year and 4 is a 4-digit
year). Replace the m4 position with a separator character.
For example, to print the date with the form yy-mm-dd,
specify &DATE=(YMD-). For December 31, 1999, the date
would appear as “99-12-31”.
The field for this form of &DATE requires 8 bytes for a 2-
digit year representation and 10 bytes for a 4-digit year. The
M, D, and Y or 4 may only appear once in the mask. All four
positions must be specified.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-108 for a description.
&DATENS=(xyz)[{±}nnnn] This form of the &DATENS subparameter specifies
that the current date or date with offset is to appear in the
report record in the form 'xyz', where x, y, and z indicate the
order in which the month, day, and year are to appear and
whether the year is to appear as two or four digits. For x, y,
and z, use M to represent the month (01-12), D to represent
the day (01-31), Y to represent the last two digits of the year
(for example, 02), or 4 to represent the four digits of the year
(for example, 2002). M, D, and Y or 4 can each be specified
only once.
For example, &DATENS=(DMY) would produce a date of
the form 'ddmmyy' which on March 29, 2002, would appear
as '290302'. &DATENS=(4MD) would produce a date of the

2–108 Syncsort MFX Programmer’s Guide

 OUTFIL

form 'yyyymmdd' which on March 29, 2002, would appear as

'20020329'. x, y, and z must be specified.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-108 for a description.
&YDDD=(m1m2m3)[{±}nnnn] This form of the &YDDD subparameter specifies
that the current date or date with offset is to appear in the
report record in the form of a year and day. You can specify
the position of the year and day, specify a separator charac-
ter, and choose between 2-digit and 4-digit year representa-
tion. The positions m1 through m3 represent masks used to
format the date. To specify the position of the year and day,
replace the m1 and m2 positions (in either position) with D
for day (001-366) and either Y or 4 for the year (where Y is a
2-digit year and 4 is a 4-digit year). Replace the m3 position
with a separator character.
For example, to print the date in the form yyyy/ddd, specify
&YDDD=(4D/). For March 29, 2005, the date would appear
as 2005/088.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-108 for a description.
&YDDDNS=(m1m2)[{±}nnnn] This form of the &YDDDNS subparameter speci-
fies that the current date or date with offset is to appear in
the report record in the form of a year and day. You can
specify the position of the year and day and choose between
2-digit and 4-digit year representation. The positions m1
and m2 represent masks used to format the date. To specify
the position of the year and day, replace the m1 and m2 posi-
tions (in either position) with D for day (001-366) and either
Y or 4 for the year (where Y is a 2-digit year and 4 is a 4-
digit year).
For example, to print the date in the form dddyy, specify
&YDDDNS=(DY). For March 29, 2005, the date would
appear as 08805.
Optionally, you can create an offset of the current date. See
“&DATE [{±}nnnn]” on page 2-108 for a description.
&TIME The &TIME subparameter specifies the current time of day
and requires 8 bytes to display hh:mm:ss, where hh is in 24-
hour format.
&TIME=(hp) This form of the &TIME subparameter generates the cur-
rent time of day and controls the formatting of the time. You
can print the time in 24-hour or 12-hour formats and specify

Syncsort MFX Programmer’s Guide 2–109

OUTFIL

the separator character between the hours, minutes, and

seconds.
The format for 24-hour time is hhpmmpss, where hh rep-
resents the hour (00-23), mm represents minutes (00-59), ss
represents seconds (00-59), and p represents the separator
character as specified by p in the &TIME=hp subparameter.
The format for 12-hour time is hhpmmpss nn, where hh rep-
resents the hour (01-12), mm represents minutes (00-59), ss
represents seconds (00-59), and p represents the separator
character as specified by p in the &TIME=hp subparameter.
The nn is “am” or “pm” as appropriate.
To select 12-hour mode specify h as 12; to select 24-hour
mode specify h as 24. The p specification represents the
character to use as a separator.
For example, to display the time in a 12-hour format with a
period as a separator, specify &TIME=(12.). At 22:43:23
hours, the time would appear as “10.43.23 pm”.
The field for this form of the &TIME subparameter requires
8 bytes for the 24-hour format and 11 bytes for the 12-hour
format.
&TIMENS=(tt) This form of the &TIMENS subparameter specifies that the
current time is to appear in the report record in the form
'hhmmss' (24-hour time) or 'hhmmss xx' (12-hour time). If tt
is 24, the time is to appear in the form 'hhmmss' (24-hour
time) where hh represents the hour (00-23), mm represents
the minutes (00-59), and ss represents the seconds (00-59).
For example, &TIMENS=(24) would produce a time of the
form 'hhmmss' which at 08:25:13 pm would appear as
'202513'. If tt is 12, the time is to appear in the form
'hhmmss xx' (12-hour time) where hh represents the hour
(01-12), mm represents the minutes (00-59), ss represents
the seconds (00-59), and xx is either ‘am’ or ‘pm’.
For a second example, &TIMENS=(12) would produce a
time of the form 'hhmmss xx' which at 08:25:13 pm would
appear as '082513 pm'.
&PAGE The &PAGE subparameter sequentially numbers logical
pages of the output report and requires 6 bytes. It produces
a 6-digit sequential page number, right justified with lead-
ing zeros suppressed.
&PAGE=(…) This subparameter is similar to the &PAGE subparameter
except that a 15-digit page number will be provided for dis-
play with editing or conversion to another data format.

2–110 Syncsort MFX Programmer’s Guide

 OUTFIL

The following describes the &PAGE subparameters that

control format conversion or printable display with edit:
fo Use this subparameter to define the out-
put
TO=fo numeric data format of an expression.
When fo is specified, mask Mm, EDIT,
and SIGNS cannot be specified. Indicate
the desired format of the output field by
replacing fo with BI, CSF/FS, FD, FI,
PD, or ZD. See “How to Convert
Numeric Data” on page 2-157 for the
default lengths of these fields. See
“LENGTH=n Subparameter” on page 2-
178 for how this default may be
changed.
TO=fo is equivalent to fo and in general,
there is no reason to use the TO= form.
However, if you are using a data dictio-
nary symbol in your control statement,
you should use the TO=fo form to avoid
ambiguities with certain types of data
conversions. See the section “INREC,
OUTREC, OUTFIL TO Subparameter”
on page 13-23.
Mm Use the Mm subparameter to indicate
that one of the 27 MFX-provided editing
masks, M0-M26, is to be used. Replace
'm' with the mask number. For details,
see “Mm Subparameter (Editing
Masks)” on page 2-179.
EDIT=(pattern) Use the EDIT subparameter to specify
that a user-provided editing mask
should be used to format the output
fields. For details, see “EDIT Subparam-
eter” on page 2-177.
SIGNS=(s1,s2,s3,s4) Use the SIGNS subparameter to specify
the signs that will appear before or after
the edited number. For details, see
“SIGNS Subparameter” on page 2-181.
LENGTH=n Use the LENGTH subparameter to alter
the length of the output field. This is
normally determined by the number of
numeric digits and either the data for-

Syncsort MFX Programmer’s Guide 2–111

OUTFIL

mat or the edit pattern and format of

the edited field. For details, see
“LENGTH=n Subparameter” on page 2-
178.
TOTAL/TOT Use the TOTAL subparameter to specify that numeric data
are to be accumulated and totaled at the end of a report, log-
ical page, or section.
After the results are included in the appropriate trailer, the
accumulator resets to zero. TOTALs either appear in print-
able format or can be converted to BI, CSF/FS, FD, FI, PD,
or ZD formats. For more information, see the fo description
on page 2-111.
If an MFX editing mask is used for totaled data, the default
length of the output field is determined by the specified
length of the input field and its format. Internally, MFX
maintains 31 digits for all data formats, but a totaled num-
ber could be bigger than the output field length. For exam-
ple, a 1 to 4 byte BI or FI field total could exceed 10 digits, or
a 1 to 8 byte PD field total or 1 to 15 byte ZD field total could
exceed 15 digits. Thus, if your totals could be that large, you
should specify the LENGTH and/or EDIT subparameters to
override the length of the output field. The following table
indicates the length that is used.

Input Format Input Length Number of digits

(bytes) needed for default
display of printable
data

BI 1-4 10

BI 5-8 20

FI 1-4 10

FI 5-8 20

FL 4 or 8 20

FS 1-16 15

FS 17-32 31

Table 18. Output Display Lengths for TOTAL and SUBTOTAL

2–112 Syncsort MFX Programmer’s Guide

 OUTFIL

Input Format Input Length Number of digits

(bytes) needed for default
display of printable
data

PD 1-8 15

PD 9-16 31

SFF/UFF 1-15 15

SFF/UFF 16-44 31

ZD 1-15 15

ZD 16-31 31

Table 18. Output Display Lengths for TOTAL and SUBTOTAL

SUBTOTAL/SUB Use the SUBTOTAL subparameter to generate a running

total of a field at the end of a report, logical page, or section.
This subparameter functions like the TOTAL subparameter
except the accumulator does not reset to zero. SUBTOTALs
either appear in printable format or can be converted to BI,
CSF/FS, FD, FI, PD, or ZD formats. For more information,
see the fo description on page 2-111.
If an MFX editing mask is used for subtotaled data, the
default length of the output field is determined by the speci-
fied length of the input field and its format. Internally, MFX
maintains 31 digits for all numeric data formats, but a
totaled number could be bigger than the output field length.
For example, a 1 to 4 byte BI or FI field total could exceed 10
digits, or a 1 to 8 byte PD field total or 1 to 15 byte ZD field
total could exceed 15 digits. Thus, if your totals could be
that large, you should specify the LENGTH and/or EDIT
subparameters to override the length of the output field.
Table 18 on page 2-112 indicates the length that is used.
MIN Use the MIN subparameter to obtain the minimum numeric
value of an input field for all records within the report, logi-
cal page, or section. MINs either appear in printable format
or can be converted to BI, CSF/FS, FD, FI, PD, or ZD for-
mats. For more information, see the fo description on page 2-
111.

Syncsort MFX Programmer’s Guide 2–113

OUTFIL

SUBMIN Use the SUBMIN subparameter to obtain the running mini-

mum numeric value of an input field for all records within
the report up to the point of the TRAILER. SUBMINs either
appear in printable format or can be converted to BI,
CSF/FS, FD, FI, PD, or ZD formats. For more information,
see the fo description on page 2-111.
MAX Use the MAX subparameter to obtain the maximum
numeric value of an input field for all records within the
report, logical page, or section. MAX values either appear in
printable format or can be converted to BI, CSF/FS, FD, FI,
PD, or ZD formats. For more information, see the fo descrip-
tion on page 2-111.
SUBMAX Use the SUBMAX subparameter to obtain the running max-
imum numeric value of an input field for all records within
the report up to the point of the TRAILER. SUBMAX values
either appear in printable format or can be converted to BI,
CSF/FS, FD, FI, PD, or ZD formats. For more information,
see the fo description on page 2-111.
AVG Use the AVG subparameter to obtain the average numeric
value of an input field for all records within the report, logi-
cal page, or section. AVG values either appear in printable
format or can be converted to BI, CSF/FS, FD, FI, PD, or ZD
formats. For more information, see the fo description on
page 2-111.
SUBAVG Use the SUBAVG subparameter to obtain the running aver-
age numeric value of an input field for all records within the
report up to the point of the TRAILER. SUBAVG values
either appear in printable format or can be converted to BI,
CSF/FS, FD, FI, PD, or ZD formats. For more information,
see the fo description on page 2-111.
p Use the p subparameter to indicate the position of the first
byte of the numeric field.
l Use the l subparameter to indicate the length of the
numeric field. Permissible lengths are 1-8 bytes for BI or FI,
4 or 8 bytes for FL, 1-16 bytes for PD, 1-31 bytes for ZD, 1-44
bytes for SFF or UFF with a 31-digit limit, and 1-32 bytes
for CSF or FS with a 31-digit limit. To determine the length
of the output field for (SUB)MIN, (SUB)MAX, and
(SUB)AVG, see “How to Convert Numeric Data” on page 2-
157.
For the (SUB)TOTAL and (SUB)AVG functions, fields are
totaled internally as 16-byte PD fields. An overflow condi-
tion will occur if the positive or negative value of a totaled or

2–114 Syncsort MFX Programmer’s Guide

 OUTFIL

subtotaled field exceeds the value that can be represented

by such fields, and the execution will terminate with an
error message.
f Use the f subparameter to indicate the format of the
numeric field. Replace f with BI, CSF, FI, FL, FS, PD, SFF,
UFF, or ZD.
fo Use this subparameter to define the output numeric data
format
TO=fo of an expression. When fo is specified, mask Mm, EDIT, and
SIGNS cannot be specified. Indicate the desired format of
the output field by replacing fo with BI, CSF/FS, FD, FI, PD,
or ZD. See “How to Convert Numeric Data” on page 2-157
for the default lengths of these fields. See “LENGTH=n Sub-
parameter” on page 2-178 for how this default may be
changed.
TO=fo is equivalent to fo and in general, there is no reason to
use the TO= form. However, if you are using a data dictio-
nary symbol in your control statement, you should use the
TO=fo form to avoid ambiguities with certain types of data
conversions. See the section “INREC, OUTREC, OUTFIL TO
Subparameter” on page 13-23.
Mm Use the Mm subparameter to indicate that one of the 27
MFX-supplied masks (M0-M26) should be used to format a
field. Replace m with the mask number. The default is M0.
For details, refer to “LENGTH=n Subparameter” on page 2-
178.
EDIT=(pattern) Use the EDIT=(pattern) subparameter to indicate that a
user-provided editing mask should be used to format a field.
For details, see “EDIT Subparameter” on page 2-177.
SIGNS=(...) Use the SIGNS subparameter to specify leading and/or
trailing signs that will appear before or after the edited
number. For details, refer to “SIGNS Subparameter” on
page 2-181.
LENGTH=(n) Use the LENGTH subparameter to alter the length of a
field determined by the edit pattern and the internal field
format. For details, refer to “LENGTH=n Subparameter” on
page 2-178.
COUNT Use the COUNT subparameter to obtain a count of the
number of records in either the entire report or a specific
part of the report. In a TRAILER1, this field will contain a
count of the total number of data records in the report. In a
TRAILER2, it will contain a count of the number of data

Syncsort MFX Programmer’s Guide 2–115

OUTFIL

records on each page. In a TRAILER3, it will contain a

count of the number of data records in each section. The
count will be the number of data records before any multi-
line OUTREC processing has been done. This number will
be a right-justified 8-digit field with leading zeros sup-
pressed. The maximum value is 99999999.
 
COUNT  + nnn =(...) This subparameter is identical to the COUNT subparame-
-
ter except that a 15-digit count will be produced for display
with editing or conversion to another data format. If the +/-
nnn subparameter is specified, the nnn value will be added
or subtracted from the count before display or conversion.
Only 3 digits may be specified for nnn.
The following sections describe the COUNT subparameters
that control format conversion or printable display with
edit:
• The fo subparameter description on page 2-111.
• “EDIT Subparameter” on page 2-177
• “Mm Subparameter (Editing Masks)” on page 2-179
• “LENGTH=n Subparameter” on page 2-178
• “SIGNS Subparameter” on page 2-181.
COUNT15 This subparameter is identical to the COUNT subparame-
ter except for the allowable size of the count number. For
COUNT15 the number will be a right-justified 15-digit field
with leading zeros suppressed. The maximum value is
999999999999999.
SUBCOUNT Use the SUBCOUNT subparameter to obtain a running, or
cumulative, count of the number of records throughout a
report. In a TRAILER1, this field will contain a count of the
total number of data records in the report. In a TRAILER2,
it will contain a cumulative count of the number of data
records on a page-by-page basis. In a TRAILER3, it will con-
tain a cumulative count of the number of data records on a
section-by-section basis. The count will be the number of
data records before any multiline OUTREC processing has
been done. This number will be a right-justified, 8-digit field
with leading zeros suppressed. The maximum value is
99999999.
SUBCOUNT=(...) This subparameter is identical to the SUBCOUNT subpa-
rameter except that a 15-digit count will be produced for
display with editing or conversion to another data format.

2–116 Syncsort MFX Programmer’s Guide

 OUTFIL

The following sections describe the SUBCOUNT subparam-

eters that control format conversion or printable display
with edit:
• The fo subparameter description on page 2-111.
• “EDIT Subparameter” on page 2-177
• “Mm Subparameter (Editing Masks)” on page 2-179
• “LENGTH=n Subparameter” on page 2-178
• “SIGNS Subparameter” on page 2-181.
SUBCOUNT15 This subparameter is identical to the SUBCOUNT subpa-
rameter except for the allowable size of the count number.
For SUBCOUNT15 the number will be a right-justified 15-
digit field with leading zeros suppressed. The maximum
value is 999999999999999.

Rules for Specifying TRAILER Subparameters

Observe the following guidelines when you specify TRAILER subparameters:
• Separate fields with commas, except for /, where commas are optional.
• Enclose literals in single quotes.
• Specify blank fields of n bytes as nX.
• If an MFX editing mask is used for totaled or subtotaled data (either by
specification or by default), the length of the generated pattern will be
determined based on the information provided in Table 18 on page 2-112,
regardless of the actual length of the field being totaled or subtotaled. Use the
LENGTH subparameter to override the length of the pattern.
• For fixed-length records, trailers specified with fewer blanks than the logical
record length (LRECL) of the output record are automatically padded on the
right with blanks.
• If a trailer length exceeds the logical record length (LRECL) of the output
record, MFX will issue the WER116A error message. If you do not wish to
shorten the trailer, you can lengthen the record. For fixed-length output, use
the OUTREC control statement or the OUTREC parameter to expand the
output record length so that it is at least as long as the longest trailer (and
header). For example, if the longest trailer is 115 characters and the output
record length is 80 bytes, use the OUTREC control statement or the OUTREC
parameter to insert a blank in position 115 of the output record. This will cause
bytes 81 through 115 to be padded with blanks. For variable-length records, it is
easiest to just change the LRECL on the output data set DD statement to
match the length of the longest trailer (and header).
• TRAILERn may not be used with the IFTRAIL parameter.

Syncsort MFX Programmer’s Guide 2–117

OUTFIL

SECTIONS Parameter (Optional)

The SECTIONS parameter allows the output report to be divided into sections.
The format of the SECTIONS parameter is illustrated below.

SECTIONS=(field1[,field2]...)

Each field is specified as follows:

p,l [,subparameter1] [,subparameter2] ...

Figure 59. SECTIONS Parameter Format

The SECTIONS parameter identifies the control field(s) that determine or control
section breaks. More than one control field can be specified to subdivide a report
within sections. However, if more than one control field is specified, the
specifications must be made in major to minor order. A major control field break
causes all minor control fields to break at the same time.
Each control field is identified by its position p and length l.
p The position value indicates the first byte of the field relative to the
beginning of the record after processing by an E15/E32 exit, the INREC
control statement, the OUTREC control statement, and an E35 exit, if
specified, but before processing by the OUTREC parameter and other
report writing parameters of the OUTFIL control statement, if specified.
l The length value indicates the length of the field. The length must be an
integer number of bytes and cannot exceed 256 bytes.
For each control field, one or more of the following subparameters may be specified:
SKIP, HEADER3, or TRAILER3. The SECTIONS subparameters are described
below.

P 
,SKIP =  nL  [,TRAILER3=(...)] [,HEADER3=(...)] [,PAGEHEAD]
 

Figure 60. SECTIONS Subparameter

SKIP The SKIP subparameter specifies the amount of spacing that
should occur after a section is completed. This spacing will
follow immediately after the last TRAILER3 for that section,
if specified. SKIP=nL specifies that the next line of the report
will appear after n number of blank lines, with n being
between 0 and 255. SKIP=P specifies a page break following
the completion of a section.
HEADER3 The HEADER3 subparameter specifies a section header or
title that will appear at the start of each new section. The

2–118 Syncsort MFX Programmer’s Guide

 OUTFIL

HEADER3 format is identical to the format of the HEAD-

ER1/HEADER2 parameters. (See HEADER1/ HEADER2
Parameters for details.)
TRAILER3 The TRAILER3 subparameter specifies a section trailer that
will appear at the end of each section. The TRAILER3 format
is identical to the format of the TRAILER1/TRAILER2
parameters. (See TRAILER1/ TRAILER2 Parameters for
details.)
PAGEHEAD The PAGEHEAD subparameter may be specified in conjunc-
tion with the HEADER3 subparameter. The PAGEHEAD sub-
parameter specifies that the HEADER3 appear at the top of
each page following any HEADER2, as well as at the start of
each new section. PAGEHEAD is ignored if no HEADER3 is
specified.
A control field may be specified without any subparameters. This allows multiple
non-contiguous control fields to be specified for each SECTIONS break field.
SECTIONS may not be used with the IFTRAIL parameter.

LINES Parameter (Optional)

Use the LINES parameter to define the logical pages constituting a report. The
pages can be defined in three ways:
• Using the carriage control characters automatically supplied by MFX
• Using ANSI control characters supplied by the user
• Using a combination of the above two methods.
Regardless of which method is selected, the number of lines defining a logical page
must be equal to or greater than the total number of lines, including blank lines,
required for all HEADER2, HEADER3, TRAILER2, and TRAILER3 entries plus at
least one record. If multiline OUTREC is used, all lines produced from each input
record will be written to the same logical page.
The format of the LINES parameter is illustrated below:

n 
 
LINES =  ANSI 
 (ANSI,n) 
 

Figure 61. LINES Parameter Format

LINES=n
If LINES=n is specified, paging is automatic and carriage control characters are
added to the beginning of each record by MFX. Because MFX requires one byte for

Syncsort MFX Programmer’s Guide 2–119

OUTFIL

a control character, the LRECL specified in the SORTOUT, SORTOFx, or

SORTOFxx DD statement must be one byte longer than the number of bytes
specified for the output record length.
Specify n as a value from 1 to 255. If report writing parameters are specified for the
file(s) (e.g., HEADERs, TRAILERs, SECTIONS), the default is LINES=60.
The LINES=n specification works in conjunction with any HEADERs and
TRAILERs you have specified as follows:
• HEADER1, if specified, prints as a preface to the report. Its page is not
numbered.
• An automatic page break occurs after HEADER1. Every nth line after the
completion of HEADER1 will signal the start of a new page.
• A HEADER2 entry, if present, is the first line(s) on each page, followed by any
HEADER3 entries that might be triggered either by control breaks or by
PAGEHEAD specifications in the SECTIONS parameter. HEADER2 is part of
the logical page.
• A HEADER3 entry, if present, is part of a section of the report. It prints as a
header for the separate report sections. HEADER3s appear in major to minor
order according to the order of their associated sections.
• If PAGEHEAD is specified, HEADER3 prints immediately below HEADER2, if
specified, or at the top of the page if a HEADER2 is not specified. A HEADER3
will not print near the end of a page if there is not sufficient room on that page
for at least one data record and a TRAILER2, if specified.
• A TRAILER3 entry, if present, is part of a section of the report. It prints as a
conclusion or summary for the separate report sections. TRAILER3s will
appear in major to minor order according to the order of their associated
sections.
• A TRAILER2 entry, if present, will be the last line(s) on the logical page,
preceded by any TRAILER3s triggered by coincidentally occurring control
breaks. TRAILER2 is part of the logical page.
• TRAILER1 will be the last page of the entire report. Its page is not numbered.
Therefore, when LINES=n is specified, all HEADER2, HEADER3, TRAILER2, and
TRAILER3 entries will be included as part of n (the total number of lines in a
logical page) and will print as described above.
LINES=ANSI
If LINES=ANSI is specified, user-provided ANSI control characters define the
logical pages. The first byte of each output record must contain an ANSI control
character (inserted, for example, by an E35 program) which is valid for the
specified output device type. For example, inserting a ‘0’ in byte 1 of the output
records produces double-spaced records.

2–120 Syncsort MFX Programmer’s Guide

 OUTFIL

The ANSI control characters which can be used with the LINES=ANSI
specification are summarized in the ANSI Control Character Chart below.
If printed output is requested, the ANSI control characters do not print as part of
the output record. If, however, the report is routed to a disk or tape device, the
control characters are included in the output data.
The LINES=ANSI specification works in conjunction with any HEADERs or
TRAILERs you have specified. If you specify HEADER2, the ANSI specification
affects this header as follows:
• After HEADER1 is output, the first logical page begins with the first line of
HEADER2.
• A logical page ends when data with a ‘1’ in the first byte are encountered. The
printing of a data record beginning with a ‘1’ is delayed until after TRAILER2
and HEADER2, if specified, are output. When record printing resumes, this
delayed record will be modified to have a control character ‘+’, which causes it to
print over the last line of HEADER2 (or HEADER3, if HEADER3 appears at
the top of the page). To prevent the data record from printing over a text line of
a header, the header should end with at least one blank line, specified by a slash
(/).
• To print HEADER2 at the top of a new physical page, the HEADER2's first line
should begin with a ‘1’.
• Because you are in complete control of the paging with LINES=ANSI, you can
permit HEADER2 to appear between variable numbers of printed records.
LINES=(ANSI,n)
If LINES=(ANSI,n) is specified, ANSI control characters govern vertical control,
and the ‘n’ specification provides additional automatic paging. Added flexibility is
provided because you can elect to double or triple space the output and still use
automatic paging.
When MFX encounters a data record with a ‘1’ in the first byte, MFX begins a new
logical page. If no data record begins with a ‘1’ but the next data record would cause
the number of lines on the page to exceed n, MFX treats the record as if it began
with a ‘1’ and begins a new page.
Refer to the LINES=ANSI section on page 2-120 for information on using a
HEADER2 with ANSI control characters.
The IFTRAIL parameter may not be used with LINES=n, LINES=ANSI, or
LINES=(ANSI,n).
Multiline OUTREC may not be used with LINES=ANSI or LINES=(ANSI,n).

Syncsort MFX Programmer’s Guide 2–121

OUTFIL

Valid ANSI Control Characters

The following chart lists the ANSI control characters accepted by MFX.

Code Interpretation Code Interpretation

blank Space one line before printing 6 Skip to channel 6 before printing

0 Space two lines before printing 7 Skip to channel 7 before printing

- Space three lines before printing 8 Skip to channel 8 before printing

+ Suppress space before printing 9 Skip to channel 9 before printing

1 Skip to channel 1 before printing A Skip to channel 10 before printing

2 Skip to channel 2 before printing B Skip to channel 11 before printing

3 Skip to channel 3 before printing C Skip to channel 12 before printing

4 Skip to channel 4 before printing V Select stacker 1

5 Skip to channel 5 before printing W Select stacker 2

Table 19. ANSI Control Character Chart

IFTRAIL Parameter (Optional)

IFTRAIL=(TRLID=(conditions),TRLUPD=(field1[,field2]...)[,HD=YES])

Figure 62. IFTRAIL Parameter Format

The IFTRAIL parameter is used to identify an existing trailer record in the input
data for an OUTFIL group and update any count or total fields in the record. This
parameter is useful when an input file is being altered in the current application by
adding or deleting records, or by modifying data fields used to produce count and
total fields in the trailer record. Using IFTRAIL lets you update the count and total
fields to reflect those changes. The updated count and total fields will reflect the
input data to OUTFIL processing.
The trailer record is identified by using the TRLID subparameter, and the updates
are specified in the TRLUPD subparameter. Since the trailer record is not a data
record, no other OUTFIL processing, such as INCLUDE/OMIT or OUTREC
parameter processing, will be performed on it. The trailer record will also not be
used for the updated count or total values. You may optionally identify the first
record passed to the OUTFIL group as a header record by using the HD=YES
parameter. A header record will similarly not be subject to other OUTFIL
processing, nor will it be used for the updated count or total values in the trailer
record.

2–122 Syncsort MFX Programmer’s Guide

 OUTFIL

IFTRAIL may not be used with any of the following OUTFIL operands: HEADERn,
TRAILERn, SECTIONS, CONVERT, VTOF, FTOV, LINES, NODETAIL, REPEAT,
SPLIT, SPLITBY, SPLIT1R, VLFILL, VLTRAIL, or VLTRIM.
TRLID=(cond) The TRLID subparameter specifies the condition that is used to
identify the trailer record in the input data to an OUTFIL group. The condi-
tion is specified in the same manner as for the OUTFIL INCLUDE parame-
ter (see p. 2-89), which is based on the INCLUDE/OMIT control statements.
For example,
TRLID=(10,6,CH,EQ,C'TOTAL:'). Note that locale processing is not used for
TRLID.
The first record into the OUTFIL group for which the condition is true is
determined to be the trailer record and thus end-of-file. All succeeding
records will be ignored for the OUTFIL group.
For variable length records, the VLTESTI installation parameter and EXEC
statement parameter does not apply. Records that do not contain all TRLID
fields will bypass TRLID processing.
TRLUPD=(field1[,field2]...) TRLUPD is used to specify the count and toral
fields to be updated in the trailer record. Each field consistes of an optional
column number (c:) followed by a COUNT[{+,-}nnn]= or TOTAL/TOT= sub-
parameter like those used in the TRAILERn parameters. These are
described on p. 2-115 and 2-112, respectively. For example,
TRLUPD=(5:COUNT=(EDIT=(IITTT)),21:TOT=23,4,ZD,M1,LENGTH=6)).
The fields should not overlap the RDW in columns 1-4 for variable length
records, and they should not extend past the end of the trailer records or
they will not be included in the record. For fixed length records, the trailer
record will be truncated or padded with blanks to match the output file
record length as necessary. If a column number is not specified, the field will
begin in column 1 or directly after the previous field. Fields should not over-
lay each other and should be specified in ascending column order.
The values used for the COUNT and TOTAL fields derive from the original
OUTFIL group input records and do not include the trailer record itself.
HD=YES HD=YES is used to identify the first OUTFIL group input record as a
header record. Normal OUTFIL processing for parameters such as
INCLUDE or OUTREC will not be applied to the header record, and the
record will not be used to determine COUNT and TOTAL values for the
trailer record, if one is found. TRLID processing will not apply to the header
record.

NODETAIL Parameter (Optional)

The NODETAIL parameter instructs the SortWriter facility to generate an output
report consisting only of header and trailer entries. Data records are not included
in the output report when this parameter is specified.

Syncsort MFX Programmer’s Guide 2–123

OUTFIL

Thus, for example, it is possible to generate a report with section trailers

containing totals and record counts without printing any data records.
NODETAIL may not be used with the IFTRAIL parameter.

REMOVECC Parameter (Optional)

The REMOVECC parameter generates reports that do not include ANSI carriage
control characters that specify printer actions (for example, skipping a line or
ejecting a page). The REMOVECC parameter omits the carriage control character
from all of the report records. REMOVECC simplifies the removal of printer
controls when output is to be displayed online or written to a list data set rather
than a printout. When REMOVECC is used, the LRECL does not require an extra
byte for the carriage control character, and the RECFM does not require the ‘A’ (for
ANSI); thus you would specify FB, not FBA.

NOTMTOFL Parameter (Optional)

 RC0 
 
NOTMTOFL=  RC4 
 RC16 
 

Figure 63. NOTMTOFL Parameter Format

The NOTMTOFL parameter specifies the action to be taken when any non-
SORTOUT OUTFIL data set contains at least one data record. NOTMTOFL will be
ignored for a BetterGener application.
RC0 The default instructs MFX to issue a return code of 0 if not overridden by a
higher return code set for another reason.
RC4 Instructs MFX to issue a WER495I warning message and continue process-
ing. A return code of 4 will be issued if not overridden by a higher return
code set for another reason.
RC16 Instructs MFX to issue a WER495A message and terminate processing with
a return code of 16.

NULLOFL Parameter (Optional)

 RC0 
 
NULLOFL=  RC4 
 RC16 
 

Figure 64. NULLOFL Parameter Format

2–124 Syncsort MFX Programmer’s Guide

 OUTFIL

The NULLOFL parameter specifies the action to be taken when any non-
SORTOUT OUTFIL data set contains no data records. NULLOFL is ignored in a
BetterGener application.
RC0 The delivered default instructs MFX to issue a return code of 0 if not over-
ridden by a higher return code set for another reason.
RC4 Instructs MFX to issue a WER461I warning message and continue process-
ing. A return code of 4 will be issued if not overridden by a higher return
code set for another reason.
RC16 Instructs MFX to issue a WER461A message and to terminate processing
with a return code of 16.

OUTPUT Parameter (Optional)

The OUTPUT parameter specifies that the OUTFIL data set will be written in a
PDF, HTML or RTF format. The corresponding OUTFIL DD must define an HFS
data set with the PATH, PATHOPTS and PATHMODE parameters. The data sets
created can then be downloaded or e-mailed (using the OUTPUT EMAIL
subparameter) to a platform that supports the viewing of these formats. If the file
is downloaded, it must be downloaded as a binary file.
If the RECFM associated with the data set includes “A” for ANSI control
characters, either due to OUTFIL report writing or because it was copied from the
input RECFM, the only printer control characters reflected in the output will be
blank, 0, - and 1. All other characters will be interpreted as blank, i.e. a new line.
For HTML, whenever a ’1’ ANSI control character is encountered, a blank line will
be generated before the record is written, except for the very first record.
The following DD statements can be added if you want to receive informational
messages from the Java environment which is used to process PDF, RTF and
HTML data sets.
//STDOUT DD SYSOUT=*
//STDERR DD SYSOUT=*
An example illustrating the use of the OUTPUT parameter appears in “Generating
Several Output Files with Different Information” on page 3-63.
The format of the OUTPUT subparameters is illustrated on the next page.

Syncsort MFX Programmer’s Guide 2–125

OUTFIL

PDF
HTML
RTF

,PORTRAIT
,LANDSCAPE
 
 LETTER 
,PAGESIZE=  LEGAL 
 
 papersize 

,MARGINS =  LEFT =  nPT  ,RIGHT = nPT ,TOP = nPT ,BOTTOM =  nPT  
 
36PT 36PT  36PT 36PT 

,TITLE = '...'

,AUTHOR = '...'

,SUBJECT = '...'

,KEYWORDS = '...'

,APPLICATION = '...'

,OWNERPASSWORD = '...'

,USERPASSWORD = '...'

 YES 
,COPYALLOWED =  
 NO 

 YES 
,PRINTINGALLOWED =  
 NO 

 WHITE 
,BACKGROUNDCOLOR =  
 ( color parameters ) 

,FONT = ( font parameters )

,FONTHn = ( font parameters )

,FONTTn = ( font parameters )

,EMAIL = ( email parameters )

Figure 65. OUTPUT Subparameters Format

PDF Specifies PDF output format. This format is the default if the
OUTPUT parameter is specified. If PDF is specified, the
HTML and RTF subparameters may not be specified.
HTML Specifies HTML output format. If HTML is specified, the PDF
and RTF subparameters may not be specified.

2–126 Syncsort MFX Programmer’s Guide

 OUTFIL

RTF Specifies RTF output format. If RTF is specified, the PDF and
HTML subparameters may not be specified.
PORTRAIT Specifies that the document should be positioned vertically.
PORTRAIT is the default if neither PORTRAIT or LAND-
SCAPE is specified.The LANDSCAPE option may not be spec-
ified if PORTRAIT is specified.
LANDSCAPE Specifies that the document should be positioned horizontally.
The PORTRAIT option may not be specified if LANDSCAPE
is specified.
PAGESIZE Specifies the page size of the output data set. LETTER (8.5” x
11”) is the default. Any of the following common paper sizes
can be specified:
_11X17, A0, A1, A10, A2, A3, A4, A5, A6, A7, A8, A9,
ARCH_A, ARCH_B, ARCH_C, ARCH_D, ARCH_E, B0, B1,
B10, B2, B3, B4, B5, B6, B7, B8, B9, CROWN_OCTAVO,
CROWN_QUARTO, DEMY_OCTAVO, DEMY_QUARTO,
EXECUTIVE, FLSA, FLSE, HALFLETTER, ID_1, ID_2,
ID_3, LARGE_CROWN_OCTAVO,
LARGE_CROWN_QUARTO, LEDGER, LEGAL, LETTER,
NOTE, PENGUIN_LARGE_PAPERBACK, PEN-
GUIN_SMALL_PAPERBACK, POSTCARD, ROYAL_OC-
TAVO,ROYAL_QUARTO, SMALL_PAPERBACK, TABLOID.
MARGINS Specifies the size of the left, right, top and bottom margins on
the page. Each value is in points and 36 is the default number
of points for each margin. One inch is equal to 72 points. The
number of points can be from 0 through 4000.
TITLE Specifies a title for the document. It can be any string up to
4095 characters. An apostrophe within the string must be
specified with double apostrophes.
AUTHOR Specifies the author of the document. It can any string up to
4095 characters. An apostrophe within the string must be
specified with double apostrophes.
SUBJECT Specifies the subject of the document. It can be any string up
to 4095 characters. An apostrophe within the string must be
specified with double apostrophes.
KEYWORDS Specifies keywords associated with the document. They can
be specified as a string of up to 4095 characters. An apostro-
phe within the string must be specified with double apostro-
phes.
APPLICATION Specifies the application name for the document. It can be
any string up to 4095 characters. An apostrophe within the

Syncsort MFX Programmer’s Guide 2–127

OUTFIL

string must be specified with double apostrophes. This subpa-

rameter is only applicable to PDF files.
OWNERPASSWORD Specifies the owner password of the document. It can be
any string up to 4095 characters. An apostrophe within the
string must be specified with double apostrophes. This subpa-
rameter is only applicable to PDF files.
USERPASSWORD Specifies the user password of the document. It can be any
string up to 4095 characters. An apostrophe within the string
must be specified with double apostrophes. This subparame-
ter is only applicable to PDF files.
COPYALLOWED Specifies whether permission is granted to copy the docu-
ment. This subparameter is only applicable to PDF files.
PRINTINGALLOWED Specifies whether permission is granted to print the doc-
ument. This subparameter is only applicable to PDF files.
BACKGROUNDCOLOR Specifies the background color for the document. Any
one of the following colors may be specified: BLACK, BLUE,
CYAN, DARKGRAY, GRAY, GREEN, LIGHTGRAY,
MAGENTA, ORANGE, PINK, RED, WHITE, YELLOW,
RGB=(int_red,int_green,int_blue). The RGB subparameters
create a color with the specified red, green, and blue values in
the range 0 to 255 or X’00’ to X’FF’. WHITE is the default.

FONT Specifies the characteristics of the font using the subparameters in

FONTHn Figure 66 on page 2-129. FONT applies to detail records.
FONTHn
FONTTn is used for headers and FONTTn is used for trailers, where n is a
number from 1 through 3. For example, FONTH1 applies to HEAD-
ER1.

2–128 Syncsort MFX Programmer’s Guide

 OUTFIL

 COURIER 
 
HELVETICA
FONTNAME=  
 TIMES_ROMAN 
 name of font 

,nPT
,12PT

,BOLD
,ITALIC
,BOLDITALIC

,BLACK
,( color parameters )

,SHADING = ( color parameters )

,UNDERLINE

Figure 66. FONT Subparameters Format

FONTNAME Specifies the name of the font. When creating a PDF

format file, only COURIER, HELVETICA and
TIMES_ROMAN are allowed. If you specify a font
other than one of these three for a PDF format, COU-
RIER will be used. For an HTML or RTF format file,
you can choose any font as your fontname. If the name
of the font is not a real font, the system default will be
used.
nPT Specifies the size of the font, where n can be a number
from 1 through 72.

BOLD Specifies whether to use bold and/or italics. The

default is none of these.
ITALIC
BOLDITALIC

color parameters Specifies the choice of color from the following list:
BLACK, BLUE, CYAN, DARKGRAY, GRAY, GREEN,
LIGHTGRAY, MAGENTA, ORANGE, PINK, RED,
WHITE, YELLOW, RGB=(int_red,int_green, int_blue).
The RGB subparameters create a color with the speci-
fied red, green, and blue values in the range 0 to 255 or
X’00 to X’FF’. BLACK is the default.

Syncsort MFX Programmer’s Guide 2–129

OUTFIL

SHADING Specifies the shading color. Choose one of the color

parameters listed above. The default is no shading.

UNDERLINE Specifies that underlining should be used. The default

is no underlining.
EMAIL Specifies that the output data set(s) defined by the OUTFIL
statement be e-mailed as an attachment (or attachments) to
one or more recipients. The file name of the attachment will
be the file name specified in the PATH parameter of the OUT-
FIL DD statement.
Figure 67 on page 2-130 displays the format of the EMAIL subparameters.

FROM = ′email_address′

,TO = ′email_address_list′

,TODD = ddname

,CC = ′email_address_list′

,CCDD = ddname

,BCC = ′email_address_list′

,BCCDD = ddname

,SUBJECT = ′text′

,BODY = ′text′

,REPLYTO = ′email_address_list′

,HOSTNAME = ′SMTP_server_name′

,PORT = { 25, n }

Figure 67. EMAIL Subparameters Format

FROM Specifies the e-mail address of the sender as

‘name@domain’.

TO Specifies one or more recipient e-mail addresses. The

addresses should be separated by commas or semi-
colons.

TODD Specifies a ddname defining one or more z/OS data

sets or an HFS file containing a list of e-mail
addresses. Each line of a file can contain one or more
complete e-mail addresses - i.e., an address cannot
span multiple lines in a file. Each address (including
the last one) must be followed by a semicolon or a

2–130 Syncsort MFX Programmer’s Guide

 OUTFIL

comma. Characters after the last semicolon or comma

in a line will be ignored. All lines of a file will be con-
catenated to form the address list. The address list for
TO and TODD will be combined.

CC Specifies one or more recipient e-mail addresses. The

addresses should be separated by commas or semi-
colons.

CCDD Specifies a ddname defining one or more z/OS data

sets or an HFS file containing a list of e-mail
addresses. Each line of a file can contain one or more
complete e-mail addresses - i.e., an address cannot
span multiple lines in a file. Each address (including
the last one) must be followed by a semicolon or a
comma. Characters after the last semicolon or comma
in a line will be ignored. All lines of a file will be con-
catenated to form the address list. The address lists
for CC and CCDD will be combined.

BCC Specifies one or more recipient e-mail addresses. The

addresses should be separated by commas or semi-
colons.

BCCDD Specifies a ddname defining one or more z/OS data

sets or an HFS file containing a list of e-mail
addresses. Each line of a file can contain one or more
complete e-mail addresses - i.e., an address cannot
span multiple lines in a file. Each address (including
the last one) must be followed by a semicolon or a
comma. Characters after the last semicolon or comma
in a line will be ignored. All lines of a file will be con-
catenated to form the address list. The address lists
for BCC and BCCDD will be combined.

SUBJECT Specifies the text of the subject line of the e-mail.

BODY Specifies the text of the body of the e-mail.

REPLYTO Specifies one or more recipient e-mail addresses. The

addresses should be separated by commas or semi-
colons.

HOSTNAME Specifies the name or IP address of the SMTP server

that will be used to send the e-mail. This can used to
override the system default.

Syncsort MFX Programmer’s Guide 2–131

OUTFIL

PORT Specifies the TCP port number that will be used to

relay the e-mail. The default is 25.

Sample OUTFIL Control Statements

Example 1
The following example illustrates how to use the OUTFIL control statement to
define multiple output files.

OUTFIL FILES=1,OUTREC=(10:1,20,40:45,5,50:60,8),
INCLUDE=(21,2,CH,EQ,C'NY')
OUTFIL FILES=2,OUTREC=(20:1,20,50:60,8),
INCLUDE=(21,2,CH,EQ,C'MA')

Figure 68. Sample OUTFIL Control Statement

The two OUTFIL control statements illustrated above are required to create two
different output files.
• The output records in the first file (SORTOF1) contain three fields from the
input record. The first input record field begins in byte 1 and is 20 bytes long,
the second input record field begins in byte 45 and is 5 bytes long, and the third
input record field begins in byte 60 and is 8 bytes long. This file will include
only those records with ‘NY’ in bytes 21 and 22 of the input record. These three
fields will begin in bytes 10, 40, and 50 of the output record.
• The output records in the second file (SORTOF2) contain two fields from the
input record. The first input record field begins in byte 1 and is 20 bytes long,
and the second input field begins in byte 60 and is 8 bytes long. This file will
include only those records with ‘MA’ in bytes 21 and 22 of the input record.
These two fields will begin in bytes 20 and 50 of the output record.
Example 2

OUTFIL FILES=(01,02,03),OUTREC=(1:1,40,50:41,40)

Figure 69. Sample OUTFIL Control Statement

This OUTFIL control statement creates three identically formatted output files:
SORTOF01, SORTOF02, and SORTOF03. These files may be written to the same
output device or to three different output devices.
• The output records contain two input record fields. The first input record field
begins in column 1. This field began in position 1 before OUTREC processing
and is 40 bytes long. The second input record field begins in column 50. This
field began in position 41 before OUTREC processing and is 40 bytes long. The
two fields will begin in positions 1 and 50 after OUTREC has been processed.

2–132 Syncsort MFX Programmer’s Guide

 OUTFIL

Example 3

OUTFIL FTOV,VLTRIM=C'*',OUTREC=(1,7,9:8,8)

Figure 70. Sample OUTFIL Control Statement with FTOV and VLTRIM

This OUTFIL control statement uses FTOV to convert fixed-length records to

variable-length records and VLTRIM to remove the specified type of trailing bytes
(in this case, asterisks).
The control statement would produce the following output:

Input Output Record Length

Records Records (with 4-byte RDW)

RECORD1ABC***** RECORD1 ABC 15

RECORD2ABCDEF** RECORD2 ABCDEF 18

RECORD3ABC****Z RECORD3 ABC****Z 20

Comprehensive examples illustrating the SortWriter facility and the multiple

output capability of the OUTFIL control statement are provided in “Chapter 3,
How to Use the MFX Data Utility Features”.

Syncsort MFX Programmer’s Guide 2–133

OUTREC

OUTREC Control Statement

The OUTREC control statement reformats the output records. Use the OUTREC
control statement to accomplish the following tasks:
• Delete or repeat segments of the input records.
• Insert character strings between data fields.
• Insert binary zeros.
• Create a sequence number field.
• Convert numeric data to printable format or to another numeric data format.
• Perform arithmetic operations (multiplication, division, modulus, addition,
subtraction) and minimum and maximum functions with numeric fields and
constants. This “horizontal arithmetic” ability complements the “vertical
arithmetic” already available with SUM, DUPKEYS, OUTFIL TOTAL, MIN,
MAX, and AVG.
• Convert data to printable hexadecimal format.
• Translate the case of EBCDIC letters from uppercase to lowercase or lowercase
to uppercase, or translate a field based on an ALTSEQ table in effect.
• Select, realign, and reorder data fields.
• Convert a variable-length record input file to a fixed-length record output file.
• Conditionally reformat records.
• Reformat only selected portions of records.
• Find and replace character or hexadecimal input constants anywhere in your
records with character, hexadecimal, or null output constants.
• Extract variable-position and variable-length fields from records and place
them into fixed-length parsed fields. These parsed fields can then be used in
any FIELDS/BUILD/OVERLAY function in which a standard p,l fixed-length
field can be used.
• Insert the current date, current time, or current date with an offset.
• Convert a field with a Julian date to a Gregorian date.
• Convert a field with a Gregorian date to a Julian date.
• Add or subtract units of days to or from an input record date field and create an
output record date field in the same format with the same length.
• Compute the interval between two date values.
The OUTREC parameter of the OUTFIL control statement can also be used to
accomplish any of the above tasks. The INREC control statement can also be used
to accomplish any of the above tasks except for converting a variable-length record

2–134 Syncsort MFX Programmer’s Guide

 OUTREC

file to a fixed-length record file. The INREC control statement also supports the
&MULTIINDD subparameter, which is used to identify the input record’s origin
when using the MULTIIN PARM.
Consider these guidelines when deciding whether to use the INREC control
statement, the OUTREC control statement, or the OUTREC parameter of the
OUTFIL control statement:
• Use the INREC control statement to delete irrelevant data fields, reformat
numeric fields to a shorter length, or combine numeric fields with arithmetic
operations and functions. Reducing the size of the input records before they are
sorted or merged usually improves performance.
• Use either the OUTREC control statement or the OUTREC parameter of the
OUTFIL control statement to expand the data record, create new numeric
fields, realign data fields, convert and edit numeric data, and change from
variable-length format to fixed-length format when you are creating one output
file.
• Use the OUTREC control statement when you are creating multiple output files
with the same output record formatting.
• Use the OUTREC parameter of the OUTFIL control statement when you are
creating multiple output files with different output record formatting.
• Use the OUTREC control statement if you need to convert a numeric field to
printable format so it can be displayed in an OUTFIL header.
• Use the OUTREC parameter of the OUTFIL control statement when an E35
exit must process the records first.
• Use the OUTREC parameter of the OUTFIL control statement when you
specify the TOTAL and/or SUBTOTAL subparameters of the TRAILER
parameter so that the accumulator(s) can sum numeric fields before they have
been converted to readable format and edited.
• Use the OUTREC parameter of the OUTFIL control statement if you want to
use the VLFILL parameter or the n/ subparameter, which are not available on
the OUTREC or INREC control statements; they can only be used with the
OUTREC parameter of the OUTFIL control statement. For a description of the
n/ subparameter, see 2-94; for the VLFILL parameter, see page 2-95.

OUTREC Control Statement Format

The format for the OUTREC control statement is illustrated below.

Syncsort MFX Programmer’s Guide 2–135

OUTREC

 
 [ PARSE=(subparm), ]  FIELDS =(fields) ,CONVERT 
 
  BUILD  ,VTOF 
OUTREC  
 IFTHEN = ( subparm ) [ ( ,IFTHEN=(subparm), ) … ] [ ,IFOUTLEN = n ] 
 [ PARSE=(subparm), ] OVERLAY=(fields) 
 
 FINDREP = ( subparm ) 

fields can be specified as follows:

 p,l [,subparm] 
 %pp [,subparm] 
 
 [n] X 
 [n] X'hhhh...hh' 
 [n] C'literal string' 
 
 [n] Z 
 'date field' 
 'time field' 
 
 1  1   (p,h)  
 SEQNUM,1,f ,START=  ---  ,INCR=  ---  ,RESTART=   
 n  i  ( %pp )  
 
 DATEADD=(datefield,number,unit) 
 DATEDIFF=(datefield 1 ,datefield 2 ,unit) 
 
[c:]  p 1 ,l 1 ,f 1yxx ,DATEDIFF,p 2 ,l 2 ,f 2yxx 
 
  ADDDAYS  
  ADDMONS  
  ADDYEARS   TOGREG= f o [ ( c ) ]  
 p,l,f yxx ,  SUBDAYS  ,numeric_field,   
    TOJUL= f o [ ( c ) ]  
  SUBMONS  
  SUBYEARS  
 NEXTDday  
  
  PREVDday  
  LASTDAYW   TOGREG= f o [ ( c ) ]  
 p,l,f yxx,  LASTDAYM  ,  
    TOJUL= f o [ ( c ) ]  
 LASTDAYQ 
 
  LASTDAYY  
 
 

Figure 71. OUTREC Control Statement Format

2–136 Syncsort MFX Programmer’s Guide

 OUTREC

PARSE Parameter
The PARSE parameter is used to extract variable-position and variable-length
fields from records and place the resultant data into fixed-length parsed fields. See
“PARSE Parameter (Optional)” on page 2-209.

IFTHEN Parameter
The IFTHEN parameter is used to conditionally reformat records. See “IFTHEN
Parameter (Optional)” on page 2-198.

OVERLAY Parameter
The OVERLAY parameter is used to reformat only selected portions of records. See
“OVERLAY Parameter (Optional)” on page 2-208.

FINDREP Parameter
The FINDREP parameter provides the ability to find and replace one or more
constants in a record. A constant to be searched for can be specified as a character
or hexadecimal string and its replacement constant can be either a character,
hexadecimal or null string. See “FINDREP Parameter (Optional)” on page 2-193 for
details.

FIELDS/BUILD Parameter
The FIELDS parameter specifies fields to be included in the output record. (BUILD
is an alias for FIELDS.)
There are three main types of fields:
• Data fields, represented by either p,l[,subparameters] for fixed fields or
%pp[,subparameters] for parsed fields
• Literal fields, to insert run-time date and time constants, character strings,
hexadecimal strings and strings of binary zeros
• Function fields, to insert sequence numbers, add or subtract values from date fields or
compute the difference between two date fields

Data field specification is defined in “Data Fields (p,l) or (%pp) Subparameters” on

page 2-137. For the specification of literal fields, see “Literal Fields
Subparameters” on page 2-166. For the specification of function fields, see
“Function Field Subparameters” on page 2-170.

Data Fields (p,l) or (%pp) Subparameters

Use the FIELDS subparameters to accomplish these tasks:
• Specify the column in which a field should begin.
• Specify halfword, fullword, or doubleword alignment.
• Convert a numeric field to a printable format with editing capabilities.

Syncsort MFX Programmer’s Guide 2–137

OUTREC

• Convert numeric data to another numeric data format.

• Perform minimum and maximum functions and arithmetic operations
(multiplication, division, modulus, addition, subtraction) with numeric fields
and constants.
• Change an input field to a replacement value in the reformatted output record
if the input field equals a search constant. The replacement value can be a
constant or another field from the input record.
• Convert a field to its printable hexadecimal representation.
• Left-justify, right-justify or “squeeze” (remove additional blanks) the data in a
field.
• Create a variable-length field from justified or squeezed fields.
• Change the case of EBCDIC letters from lowercase to uppercase or vice-versa,
translate ASCII characters to EBCDIC ones or vice-versa, or transform data to
printable hexadecimal (0-9 or A-F) or binary (0 or 1), or vice-versa, or translate
data based on an alternative collating sequence (ALTSEQ) table in effect.

• Convert date data.

• Convert a 2-digit year field to a 4-digit year field.

• Convert a full-date field to a printable field with separators.
• Convert any full-date field to a Gregorian or Julian date field.
The figure below illustrates how the FIELDS subparameters should be specified
and describes their functions. For information on the EDIT, LENGTH, Mm, and
SIGNS subparameters, see “How to Convert Numeric Data” on page 2-157.

2–138 Syncsort MFX Programmer’s Guide

 OUTREC

 
  ,f o [,LENGTH=n]  
   
  ,TO = f o [,LENGTH=n]  
   
 expression  , Mm   
 
 , M0   
    [,SIGNS=(...)] [,LENGTH=n]  
 
 , EDIT=(...)   
     
 
  ,a  
   
  ,CHANGE=(........) [,NOMATCH=(....)]  
  ,JFY=(....)  
  ,SQZ=(....)  
    
 p,l 
  ,f yxx (c) 
  %pp   ,f P  
  y 2f  
[c:]    
  ,DT [ = ( m 1 m 2 m 3 m 4 ) ]  
 ,f yxx   
   
  ,DTNS [ = ( xyz ) ]  
 
 
  ,HEX  
   
     
   LTOU   
   UTOL   
   ATOE    
     
  p [,l]    ETOA   
  %pp   ,TRAN=  HEX  
   UNHEX   
    
   BIT  
   UNBIT   
   ALTSEQ  
    
 

Figure 72. FIELDS Subparameters Format

Each data field specified in the FIELDS parameter is identified either by its
position p and length l or by its %pp identifier for parsed fields.
p For INREC, the position value indicates the first byte of the
field relative to the beginning of the input record after E15
processing, if specified, has completed. For OUTREC, the
position value indicates the first byte of the field after both
E15 and INREC processing, if specified, have completed. If
the OUTREC parameter of the OUTFIL control statement is
used, the position value refers to the record after E35 process-
ing as well. The field must begin on a byte boundary.
l The length value indicates the length of the field. The length
must be an integer number of bytes.
%pp Identifies a fixed-length parsed field. See “PARSE Parameter
(Optional)” on page 2-209 for information on creating parsed
%pp fields.

Syncsort MFX Programmer’s Guide 2–139

OUTREC

The following describes the c: subparameter:

c: Use the c: subparameter to define the column in which the field should
begin. MFX will add the appropriate number of blanks to achieve the proper
alignment. This subparameter can be specified for all types of fields.
The term expression represents the following syntax:

 p,l,f 
 i 
 %pp,f i 
 
 +n 
 -n 
 [ ( ''expression '' [,''operator'',''expression '' ] ) ] 
 1 2 
 

Figure 73. Syntax for expression

The following describes the elements of expression:
p,l,fi This specifies the position, length, and format of an input
field. (See the description of fi below for details.)
%pp,fi Identifies a fixed-length parsed field and format. (See the
description of fi below for details.)
+n This represents a positive numerical constant of up to 31 dec-
imal digits. The + sign must be specified.
-n This represents a negative numerical constant of up to 31 dec-
imal digits. The - sign must be specified.
expression An expression defines a numeric value. The simplest forms of
an expression consist of a numeric data input field defined
either by p,l,fi or a constant defined by +n or -n. Expressions
can also be created by connecting these simple expressions
with operators, as shown in the last line of the above syntax
illustration. Parentheses may be used to change the default
precedence order of the operators. Algebraic equations can
thus be represented with an expression.
A maximum value of 31 digits is permitted at all times in
evaluating an expression. If this is exceeded, a critical error
will be issued. Similarly, an attempted division by zero will
also result in a critical error. The results of division will be
rounded down to an integer.
Once an expression has been defined, its value can either be
converted to a numeric output data format or to a printable
numeric format using editing masks. See “How to Convert
Numeric Data” on page 2-157. The default is to use the M0
editing mask to create printable output. The number of digits

2–140 Syncsort MFX Programmer’s Guide

 OUTREC

in an expression is defined to be 31 unless the expression is a

simple p,l,fi field.
The following are expressions:

+10
10,2,Y2Z
+10,ADD,10,2,Y2Z
1,4,ZD
10,2,PD
+30
1,4,ZD,ADD,10,2,PD
+30,MUL,(1,4,ZD,ADD,10,2,PD)
+30,MUL,(1,4,ZD,ADD,10,2,PD),MIN,(5,5,ZD,DIV,+100)
(+30,MUL,(1,4,ZD,ADD,10,2,PD)),MIN,(5,5,ZD,DIV,+100
)
operator Operations between two numeric fields or constants are per-
formed with operators. There are two types of operators: func-
tion operators and arithmetic operators. The following are the
function operators:
MIN Generates the minimum arithmetic value of two
specified fields.
MAX Generates the maximum arithmetic value of two
specified fields.
The following are the arithmetic operators:
MUL multiplication
DIV division
MOD modulus
ADD addition
SUB subtraction
The following rules of arithmetic precedence apply in comput-
ing an “expression”:
• Conditions within parentheses are evaluated first, from
innermost to outermost parentheses.
• The arithmetic functions of minimum and maximum
(MIN and MAX) are performed before the arithmetic
operators (MUL, DIV, MOD, ADD, SUB). Within the
arithmetic operators, multiplication (MUL), division
(DIV), and modulus (MOD) are performed before addition
(ADD) and subtraction (SUB). Operations within the
same precedence level are performed from left to right.

Syncsort MFX Programmer’s Guide 2–141

OUTREC

The result of the DIV operation is truncated (rounded

down) to an integer. The MOD operation produces an inte-
ger remainder with the sign of the dividend.
fi Use this parameter together with p,l to define the input for-
mat of a numeric field that is part or all of an expression. The
expression will then be converted to either another numeric
data format or to a printable format. In such cases, indicate
the format of the data field that is to be converted by replac-
ing fi with BI, FI, FL, PD, ZD, CSF/FS, PD0, SFF, UFF, one of
the SMF formats (DT1, DT2, DT3, TM1, TM2, TM3, and
TM4), time-of-day (TOD) formats (DC1, DC2, DC3, TC1, TC2,
TC3, TC4), extended time-of-day (ETOD) formats (DE1, DE2,
DE3, TE1, TE2, TE3, TE4), or one of the year data formats
(Y2B, Y2C, Y2D, Y2P, Y2S, Y2Z, Y2T, Y2U, Y2V, Y2W, Y2X,
Y2Y, Y4T, Y4U, Y4V, Y4W, Y4X, Y4Y).
Also use this parameter when a 2-digit packed decimal year
value is to be expanded to a 4-digit packed decimal value. In
such cases replace fi with Y2ID or Y2IP. The Y2ID and Y2IP
formats cannot be used to form complex arithmetic expres-
sions and do not allow the specification of mask (Mm), EDIT,
SIGNS, or LENGTH.
An l value indicating the length of the field must be specified
in accordance with the following allowable values:
for BI ... 1-8 inclusive
for CSF or FS ... 1-16 inclusive (15-digit limit)
for CSF or FS ... 17-32 inclusive (31-digit limit)
for FI ... 1-8 inclusive
for FL ... 4 or 8
for PD ... 1-16 inclusive
for PD0 ... 2-8 inclusive
for SFF ... 1-44 inclusive (31-digit limit)
for UFF ... 1-44 inclusive (31-digit limit)
for Y2B ... 1
for Y2C ... 2
for Y2D ... 1
for Y2ID ... 1
for Y2IP ... 2
for Y2P ... 2
for Y2S ... 2
for Y2Z ... 2
for ZD ... 1-31 inclusive
for Y2T ... 3-6 inclusive
for Y2U ... 2-3 inclusive
for Y2V ... 3-4 inclusive

2–142 Syncsort MFX Programmer’s Guide

 OUTREC

for Y2W ... 3-6 inclusive

for Y2X ... 2-3inclusive
for Y2Y ... 3-4inclusive
for Y4T ... 7 or 8
for Y4U ... 4
for Y4V ... 5
for Y4W ... 7 or 8
for Y4X ... 4
for Y4Y ... 5
Field conversion of a single p,l,fi expression with a format of
Y2x, Y2xx, Y4x does not default to the use of the M0 default
output mask. Y4x fields will be converted to printable format
and for Y2x/Y2xx fields the default will convert the 2-digit
year portion to a 4-digit 4-byte printable year. The year por-
tion of the date is converted using the century window defined
by the CENTWIN parameter. The century window is not used
for the special values, which are only expanded with charac-
ters of the proper format. However, except for Y2S, Y2x and
Y4x, fields can be used to form expressions with operators. In
this case, the default will use the M0 output mask with a
number of decimal digits determined by the terms used in the
expression. For more information, see “How to Convert
Numeric Data” on page 2-157. The specification of an output
numeric data format fo or mask Mm, EDIT, SIGNS, or
LENGTH is permitted except when using Y2S, Y2ID, and
Y2IP.
Field conversion of a single p,l,fi expression with a format of
FL will convert a hexadecimal floating point value (4-byte or
8-byte) to a signed integer in the range of
-9223372036854775808 to 9223372036854775807, with the
fractional part of the FL value dropped. A z/Architecture envi-
ronment is required before specifying FL format; otherwise,
an error message and termination of the application will
result.
The following describes the other FIELDS subparameters:
fo Use this subparameter to define the output numeric data for-
mat of an
TO=fo expression. When fo is specified, mask Mm, EDIT, and SIGNS
cannot be specified. Indicate the desired format of the output
field by replacing fo with BI, CSF/FS, FD, FI, PD, PDC, PDF,
ZD, ZDC or ZDF. The PDC format represents a PD field and
uses a C for the sign of a positive value and D for the sign of a
negative value. The ZDC format represents a ZD field and
uses a C for the sign of a positive value and D for the sign of a

Syncsort MFX Programmer’s Guide 2–143

OUTREC

negative value. PDF produces the same numerical value as

PD, but uses an F for a positive sign and D for the sign of a
negative value. ZDF produces the same numerical value as
ZD, but uses an F for a positive sign and D for the sign of a
negative value. See “How to Convert Numeric Data” on
page 2-157 for the default lengths of these fields. See
“LENGTH=n Subparameter” on page 2-178 for how this
default may be changed.
TO=fo is equivalent to fo and in general, there is no reason to
use the TO= form. However, if you are using a data dictionary
symbol in your control statement, you should use the TO=fo
form to avoid ambiguities with certain types of data conver-
sions. See the section “INREC, OUTREC, OUTFIL TO Subpa-
rameter” on page 13-23.
Mm Use the Mm subparameter to indicate that one of the 27
MFX-provided editing masks, M0-M26, is to be used. Replace
'm' with the mask number. For details, see “Mm Subparame-
ter (Editing Masks)” on page 2-179.
EDIT=(pattern) Use the EDIT subparameter to specify that a user-provided
editing mask should be used to format the output fields. For
details, see “EDIT Subparameter” on page 2-177.
SIGNS=(s1,s2,s3,s4) Use the SIGNS subparameter to specify the signs that will
appear before or after the edited number. For details, see
“SIGNS Subparameter” on page 2-181.
LENGTH=n Use the LENGTH subparameter to alter the length of the out-
put field. This is normally determined by the number of
numeric digits d and either the data format or the edit pat-
tern and format of the edited field. For details, see
“LENGTH=n Subparameter” on page 2-178.
a Use this subparameter to tell MFX how the field should be
aligned with respect to the start of the output record. Replace
a with H, F, or D to specify halfword (H), fullword (F), or dou-
bleword (D) alignment. The alignment itself actually takes
place after the column designation. It will automatically pad
any provided field with the number of bytes of binary zeros
required to achieve the specified alignment. This subparame-
ter cannot be used in conjunction with data conversion.
CHANGE=(........)/
NOMATCH=(...) Use the CHANGE subparameter to change an input field to a
replacement constant or input record field in the reformatted
output record if the input field equals a search constant. For a
complete description, see “CHANGE Subparameter” on
page 2-182.

2–144 Syncsort MFX Programmer’s Guide

 OUTREC

JFY=(...) Use the JFY subparameter to specify that an input field be

processed for left-justification or right-justification for the
output record. For left-justification, leading blank characters
are eliminated; the remaining characters are shifted left; if
necessary, blank characters are introduced to the right. For
right-justification, trailing blank characters are eliminated;
the remaining characters are shifted right; if necessary, blank
characters are introduced to the left. Options include intro-
ducing new leading and trailing nonblank characters; elimi-
nating previous leading and trailing nonblank characters;
and changing the length of the field in the output record. For
a complete description and options, see “JFY Subparameter”
on page 2-186.
SQZ=(...) Use the SQZ subparameter to specify that an input field be
processed for “left-squeezing” or “right-squeezing” for the out-
put record. SQZ includes the justification functions of the JFY
subparameter but adds elimination of all blank characters in
the input field and additional options for selecting and replac-
ing blank and nonblank characters, which include introducing
leading and trailing nonblank characters; replacing user-
specified nonblank characters with blank characters prior to
squeeze operation; replacing blank characters with user-spec-
ified nonblank characters; retaining blank characters
between paired apostrophes and paired quotes; and changing
the length of the field in the output record. For a complete
description and options, see “SQZ Subparameter” on page 2-
189.
HEX Use the HEX subparameter to convert a record field to its
hexadecimal representation. Specify this subparameter
immediately after the position p and the length l of the field
to be converted. Specify p,l,HEX for both fixed-length records
and the fixed-length portion of variable-length records. Spec-
ify p,HEX for the variable-length portion of variable-length
records. Starting in position p of the input record, for a length
of l, each byte will be converted to its hexadecimal representa-
tion. Note that in the reformatted record, the converted field
will be twice the length of the original field.

TRAN Use this subparameter to change the case of EBCDIC letters

from lowercase to uppercase or vice-versa, translate ASCII
characters to EBCDIC ones or vice-versa, transform data to
printable hexadecimal (0-9 or A-F) or binary (0 or 1), or vice-
versa, or translate data based on an alternative collating
sequence (ALTSEQ) table in effect. Specify this subparameter
immediately after the position p and the length l of the field

Syncsort MFX Programmer’s Guide 2–145

OUTREC

to be converted. Specify p,l,TRAN for both fixed-length

records and the fixed-length portion of variable-length
records. Specify p,TRAN for the variable-length portion of
variable-length records. Starting in position p of the input
record, for a length of l, each byte will be converted as per
specification.

 
 LTOU 
 UTOL 
 ATOE 
 
 ETOA 
TRAN=  HEX 
 UNHEX 
 
 BIT 
 UNBIT 
 ALTSEQ
 

Figure 74. TRAN Subparameter Format

LTOU Instructs MFX to translate EBCDIC letters in a
specified field from lowercase to uppercase.
UTOL Instructs MFX to translate EBCDIC letters in a
specified field from uppercase to lowercase.
ALTSEQ Instructs MFX to translate characters based on
the ALTSEQ table in effect.
ATOE Instructs MFX to translate characters in a speci-
fied field from ASCII to EBCDIC. The maximum
input length is 32752.
ETOA Instructs MFX to translate characters in a speci-
fied field from EBCDIC to ASCII. The maximum
input length is 32752.
HEX Instructs MFX to transform data to printable hexa-
decimal. The number of output bytes will be 2x the
number of input bytes. The maximum input length
is 16376. For example, C'B9' is transformed into
C'C2F9'.
UNHEX Instructs MFX to transform data from printable
hexadecimal to its character representation. Each
two input bytes are translated to one output byte,
so the number of output bytes will be one-half the
number of input bytes, rounded up. If the input
length is an odd number, the output will be padded
with binary zeros in the last half-byte. The maxi-
mum input length is 32752. Input bytes that are

2–146 Syncsort MFX Programmer’s Guide

 OUTREC

not in the range 0-9 or A-F will be treated as 0. For

example, C'C4FX5' is interpreted as C'C4F050' and
is transformed into C'D0&'.
BIT Instructs MFX to transform data to printable
binary. The number of output bytes will be 8x the
number of input bytes. The maximum input length
is 4094. For example, C'F12' is equivalent to
X'C6F1F2' and is transformed into
C'110001101111000111110010'.
UNBIT Instructs MFX to transform data from printable
binary to its character representation. Each eight
input bytes are translated to one output byte, so
the number of output bytes will be one-eighth the
number of input bytes, rounded up. If the input
length is not a multiple of 8, the output will be pad-
ded with binary zeros in the last byte. The maxi-
mum input length is 32752. Input bytes that are
not 0 or 1 will be treated as 0.
For example, C’1111a02111001' is interpreted as
C'1111000111001000' (or X'F1C8') and is trans-
formed into C'1H'.
For examples of OUTREC control statements that use the
TRAN subparameter, see Figure 125 on page 2-223 and Fig-
ure 126 on page 2-223.
fy2f(c) Use this subparameter together with the p,l elements to indi-
cate the conversion of a full-date field to a printable date with
separator character(s). The “c” represents the separator and
can be any character except a blank. For Y2x fields, the year
portion of the date is converted to a 4-digit year using the cen-
tury window defined by the CENTWIN parameter. The cen-
tury window is not used for the special values, which are
expanded with characters of the proper format. (See Table 20
on page 2-148.)

Syncsort MFX Programmer’s Guide 2–147

OUTREC

The following table shows what is produced if (c) is set to a “/”:

Full-Date Input Length

Date Form Output Format
Format (bytes)

Y2T yyx 3 yyyy/x

yyxx 4 yyyy/xx

yyxxx 5 yyyy/xxx

yyxxxx 6 yyyy/xx/xx

Y2U yyx 2 yyyy/x

(X'yyxs')

yyxxx 3 yyyy/xxx

(X'yyxxxs')

Y2V yyxx 3 yyyy/xx

(X'0yyxxs')

yyxxxx 4 yyyy/xx/xx

(X'0yyxxxxs')

Y2W xyy 3 x/yyyy

xxyy 4 xx/yyyy

xxxyy 5 xxx/yyyy

xxxxyy 6 xx/xx/yyyy

Y2X xyy 2 x/yyyy

(X'xyys')

xxxyy 3 xxx/yyyy

(X'xxxyys')

Y2Y xxyy 3 xx/yyyy

(X'0xxyys')

xxxxyy 4 xx/xx/yyyy

(X'0xxxxyys')

Table 20. Full-Date Field Conversions

2–148 Syncsort MFX Programmer’s Guide

 OUTREC

fy2fP Use this subparameter together with the p,l elements to indi-
cate the conversion of a full-date field to a packed decimal for-
mat. The year portion of the date is converted to a 4-digit year
using the century window defined by the CENTWIN parame-
ter. The century window is not used for the special values,
which are expanded with characters of the proper format. See
Table 21 on page 2-149.

Full-Date Input Length

Date Form Output Format*
Format (bytes)

Y2TP yyx 3 X'yyyyxC'

yyxx 4 X'0yyyyxxC'

yyxxx 5 X'yyyyxxxC'

yyxxxx 6 X'0yyyyxxxxC'

Y2UP yyx 2 X'yyyyxC'

(X'yyxs')

yyxxx 3 X'yyyyxxxC'

(X'yyxxxs')

Y2VP yyxx 3 X'0yyyyxxC'

(X'0yyxxs')

yyxxxx 4 X'0yyyyxxxxC'

(X'0yyxxxxs')

Y2WP xyy 3 X'xyyyyC'

xxyy 4 X'0xxyyyyC'

xxxyy 5 X'xxxyyyyC'

xxxxyy 6 X'0xxxxyyyyC'

Y2XP xyy 2 X'xyyyyC'

(X'xyys')

xxxyy 3 X'xxxyyyyC'

(X'xxxyys')

Table 21. (Page 1 of 2)Full-Date Field Conversions fy2fP

Syncsort MFX Programmer’s Guide 2–149

OUTREC

Full-Date Input Length

Date Form Output Format*
Format (bytes)

Y2YP xxyy 3 X'0xxyyyyC'

(X'0xxyys')

xxxxyy 4 X'0xxxxyyyyC'

(X'0xxxxyys')

Y2PP x'0yys' 2 X'yyyy'

Y2DP x'yy' 1 X'yyyy'

Y4T yyyyddd 7 yyyy/ddd

yyyymmdd 8 yyyy/mm/dd

Y4W dddyyyy 7 ddd/yyyy

mmddyyyy 8 mm/dd/yyyy

Y4U yyyyddd 4 yyyy/ddd

(X'yyyyddds')

Y4V yyyymmdd 5 yyyy/mm/dd

(X'0yyyymmdds')

Y4X dddyyyy 4 ddd/yyyy

(X'dddyyys')

Y4Y mmddyyyy 5 mm/dd/yyyy

(X'0mmddyyyys')

* 'C' is a positive sign value.

Table 21. (Page 2 of 2)Full-Date Field Conversions fy2fP

,DT [ = ( m 1 m 2 m 3 m 4 ) ]
{ f y2f ,f y4f } 
,DTNS [ = ( m1 m 2 m 3 ) ] Use this subparameter together with the p,l elements
to indicate the conversion of a full-date field to a printable
Gregorian date. The resultant field can be created with a sep-
arator character by specifying the DT subparameter or with-
out a separator by specifying the DTNS subparameter. For
Y2x fields, the year portion of the date is converted to a 4-
digit year using the century window defined by the
CENTWIN parameter. Invalid input date values will be con-
verted to all nines in the digit fields. All full-date Y2x and
YYx fields are vald input fields. (See page 2.141)
DT[=(m1m2m3m4)] This form of the subparameter converts
the date and controls the formatting of the Grego-
rian date. You can specify the position of the year,

2–150 Syncsort MFX Programmer’s Guide

 OUTREC

month, and day, specify a separator character, and

choose between 2-digit and 4-digit year representa-
tion. The positions m1 through m4 represent masks
used to format the date. To specify the position of
the month, day, and year, replace the m1, m2, and
m3 positions, in any order, with M for the month
(01-12), D for the day (01-31), and either Y or 4 for
the year (where Y is a 2-digit year and 4 is a 4-digit
year). Replace the m4 position with a separator
character. If the mask specification is omitted, a
mask of 'MDY/' will be used by default.
For example, if an input field contains a character
Julian date in the form of yyddd, then to convert to
a Gregorian date in a month, day, 4-digit year for-
mat with a / separator, specify p,l,Y2T,DT=(MD4/).
For December 31, 2007, the input field would be
'07365' and the output Gregorian date would
appear as '12/31/2007'.
The field for this form requires 8 bytes for a 2-digit
year representation and 10 bytes for a 4-digit year
representation. The M,D, and Y or 4 may only
appear once in the mask.
DTNS[=(m1m2m3)] This form of the subparameter specifies
that the full date is to be converted in the form
'm1m2m3', where m1, m2, and m3 indicate the order
in which the month, day, and year are to appear
and if the year is to appear as two or four digits.
For m1, m2, and m3, use M to represent the month
(01-12), D to represent the day (01-31), Y to repre-
sent the last two digits of the year (for example,
02), or 4 to represent the four digits of the year (for
example, 2002). If the mask specification is omit-
ted, a mask of 'MDY' will be used by default.
For example, if an input field contains a character
Julian date in the form of yyddd, then to convert to
a Gregorian date in a month, day, 4-digit year for-
mat, specify p,l,Y2T,DTNS=(MD4). For December
31, 2007, the input field would be '07365' and the
output Gregorian date would appear as '12312007'.
The field for this form requires 6 bytes for a 2-digit
year representation and 8 bytes for a 4-digit year
representation. The M, D, and Y or 4 may only
appear once in the mask.

Syncsort MFX Programmer’s Guide 2–151

OUTREC

,YD [ = ( m 1 m 2 m 3 ) ] Use this subparameter

together with the p,l elements to
f y2fg 
,YDNS [ = ( m 1 m 2 ) ] indicate the conversion of
a full-date Gregorian date field to a
printable Julian date. The resultant field can be created with
a separator character by specifying the YD subparameter or
without a separator by specifying the YDNS subparameter.
The year portion of the date is converted to a 4-digit year
using the century window defined by the CENTWIN parame-
ter. Invalid input date values will be converted to all nines in
the digit fields. See Table 22 on page 2-152 for the valid for-
mats that can be specified for the Gregorian dates.

Full Date Format Date Form Input Length

Y2T yymmdd 6

Y2V X'0yymmdds' 4

Y2W mmddyy 6

Y2Y X'0mmddyys' 4

Table 22. Valid Length l and Format f Combinations

YD[=(m1m2m3)] This form of the subparameter converts the

Gregorian date and controls the formatting of the
Julian date. You can specify the position of the year
and day, specify a separator character, and choose
between 2-digit and 4-digit year representation.
The positions m1 through m3 represent masks used
to format the date. To specify the position of the
day and year, replace the m1 and m2 positions, in
any order, with D for the day (001-366), and either
Y or 4 for the year (where Y is a 2-digit year and 4
is a 4-digit year). Replace the m3 position with a
separator character. If the mask specification is
omitted, a mask of 'DY/' will be used by default.
For example, if an input field contains a character
Gregorian date in the form of yymmdd, then to con-
vert to a Julian date in a day and 4-digit year for-
mat with a / separator, specify p,l,Y2T,YD=(D4/).
For December 31, 2007, the input field would be
'071231' and the output Julian date would appear
as '365/2007'.
The field for this form requires 6 bytes for a 2-digit
year representation and 8 bytes for a 4-digit year

2–152 Syncsort MFX Programmer’s Guide

 OUTREC

representation. The D and Y or 4 may only appear

once in the mask.
YDNS[=(m1m2)] This form of the subparameter specifies
that the Gregorian date is to be converted in the
form 'm1m2', where m1 and m2 indicate the order in
which the day and year are to appear and if the
year is to appear as two or four digits. For m1 and
m2, use D to represent the day (001-366), Y to rep-
resent the last two digits of the year (for example,
02), or 4 to represent the four digits of the year (for
example, 2002). If the mask specification is omit-
ted, a mask of 'DY' will be used by default.
For example, if an input field contains a character
Gregorian date in the form of yymmdd, then to con-
vert to a Julian date in a day and 4-digit year for-
mat, specify p,l,Y2T,YDNS=(D4). For December 31,
2007, the input field would be '071231' and the out-
put Julian date would appear as '3652007'.
The field for this form requires 5 bytes for a 2-digit
year representation and 7 bytes for a 4-digit year
representation. The D and Y or 4 may only appear
once in the mask.
Full-date year fields can be converted to other full-date year-field formats using the
TOJUL and TOGREG parameters. For printable output formats, you may insert a
separator character. Input and output formats may be either 2-digit years (Y2x
fomats) or 4-digit years (Y4x formats). When expanding a 2-digit year field to a 4-
digit year field, CENTWIN processing is used to determine the high-order yy. Full-
date year fields can also be converted to an output field that is an indicator of the
day of the week of the field using the WEEKDAY parameter. The eligible input
formats are:

Format Length Date Form

Y2T 5 C'yyddd'

Y2T 6 C'yymmdd'

Y2U 3 X'yyddds' (P'yyddd')

Y2V 4 X'0yymmdds' (P'yymmdd')

Table 23. Eligible Input Date Formats for TOJUL, TOGREG, and WEEKDAY

Syncsort MFX Programmer’s Guide 2–153

OUTREC

Format Length Date Form

Y2W 5 C'dddyy'

Y2W 6 C'mmddyy'

Y2X 3 X'dddyys' (P'dddyy')

Y2Y 4 X'0mmddyys' (P'mmddyy')

Y4T 7 C'yyyyddd'

Y4T 8 C'yyyymmdd'

Y4U 4 X'yyyyddds' (P'yyyyddd')

Y4V 5 X'0yyyymmdds' (P'yyyymmdd')

Y4W 7 C'dddyyyy'

Y4W 8 C'mmddyyyy'

Y4X 4 X'dddyyyys' (P'dddyyyy')

Y4Y 5 X'0mmddyyyys' (P'mmddyyy')

Table 23. Eligible Input Date Formats for TOJUL, TOGREG, and WEEKDAY

fiyxx,TOJUL=foyxx TOJUL converts full-date Gregorian or

Julian year fields into a Julian format full-date
field. For printable output fields, TOJUL=foyxx(c)
may be used to create an output date with c as the
separator, where c can be any character except a
blank. The eligible output foyxx formats are listed
below along with the length and format of the
fields.

For TOJUL=foyxx For TOJUL=foyxx(c)

Format Length Date Form Format Length Date Form

Y2T 5 C'yyddd' Y2T 6 C'yycddd'

Table 24. Julian Date Output Formats

2–154 Syncsort MFX Programmer’s Guide

 OUTREC

For TOJUL=foyxx For TOJUL=foyxx(c)

Format Length Date Form Format Length Date Form

Y2U 3 X'yyddds' (P'yyddd') Y2W 6 C'dddcyy'

Y2W 5 C'dddyy' Y4T 8 C'yyyycddd'

Y2X 3 X'dddyys' (P'dddyy') Y4W 7 C'dddyyyy'

Y4T 7 C'yyyyddd' Y4X 4 X'dddyyyys' (P'dddyyyy')

Y4U 4 X'yyyyddds' (P'yyyyddd') Y4W 8 C'dddcyyyy'

Table 24. Julian Date Output Formats

fiyxx,TOGREG=foyxx TOGREG converts full-date Gregorian

or Julian year fields into a Gregorian format full-
date field. For printable output fields,
TOGREG=foyxx(c) may be used to create an otuput
date with c as the separator, where c can be any
character except a blank. The eligible output foyxx
formats are listed below along with the length and
format of the fields.

For TOGREG=foyxx For TOGREG=foyxx(c)

Format Length Date Form Format Length Date Form

Y2T 6 C'yymmdd' Y2T 7 C'yycmmdd'

Y2V 4 X'0yymmdds' (P'yymmdd') Y2W 7 C'mmddcyy'

Y2W 6 C'mmddyy' Y4T 8 C'yyyymmdd'

Y2Y 3 X'0mmddyys' (P'mmddyy') Y4T 9 C'yyyycmmdd'

Y4V 5 X'0yyyymmdds' (P'yyyymmdd') Y4W 9 C'mmddcyyyy'

Y4Y 5 X'0mmddyyyys' (P'mmddyyyy') Y4W 8 C'mmddyyyy'

Table 25. Gregorian Date Output Formats

Syncsort MFX Programmer’s Guide 2–155

OUTREC

fiyxx,WEEKDAY={CHAR3,CHAR9,DIGIT1} The WEEK-

DAY parameter creates an output field that is a
printable indicator of the day of the week of the
full-date input field. There are 3 types of output
indicators that you can create. The output field
lengths for CHAR3, CHAR9, and DIGIT1 and 3, 9,
and 1 respectively. CHAR9 is right-padded with
blanks when necessary.

CHAR3 CHAR9 DIGIT1

SUN SUNDAY 1

MON MONDAY 2

TUE TUESDAY 3

WED WEDNESDAY 4

THU THURSDAY 5

FRI FRIDAY 6

SAT SATURDAY 7

Table 26. WEEKDAY Output Fields

Specifying the FIELDS Parameter for Variable-Length Records

If you are not using the CONVERT option to convert variable-length records to
fixed-length records, you must observe these rules when you specify the FIELDS
parameter for variable-length records:
• Remember to specify 4 bytes for the Record Descriptor Word in the first output
field. You can include the 4 bytes in the length value of the first field if the first
field in the original data record is also the first field specified in the FIELDS
parameter.
• To include any portion of the variable part of the input records, specify a
position value without a length value as the last entry. The only subparameters
you can specify after the position value are the HEX and TRAN conversion
subparameters. (Refer to the FIELDS subparameters sections on HEX and
TRAN on pages 2-145 and 2-146.)
• If INREC or OUTREC processing changes the output record length, the
contents of the Record Descriptor Word will be automatically revised by the
sort.

2–156 Syncsort MFX Programmer’s Guide

 OUTREC

How to Convert Numeric Data

One of the most important functions of OUTREC processing is to convert a numeric
data field or an expression to either an output numeric data format or a printable
format with editing capabilities. OUTREC processing can convert 2-digit year
fields into 4-digit year fields, as well as any 2-digit or 4-digit year full-date field into
any other full-date field, including conversion between Julian and Gregorian
formats. For details on converting 2-digit year data, see “Converting Year Data
with Century Window Processing on INREC, OUTREC, or OUTFIL OUTREC” on
page 2-162. When a single numeric field defined by p,l,fi is to be converted to a
printable format without editing, the format and length of the field determine the
length of the output field, as illustrated in the following two tables.

Data Conversion

Input Format Bytes in Input Field Resulting Digits (d)

ZD n n

PD n 2n-1

BI, FI 1 3

BI, FI 2 5

BI, FI 3 8

BI, FI 4 10

BI, FI 5 13

BI, FI 6 15

BI, FI 7 17

BI, FI 8 20

CSF or FS n=1 to 16 n (to maximum of 15, then

truncated)

CSF or FS n=17 to 32 n (to maximum of 31, then

truncated)

FL 4 or 8 20

PD0 n 2n-2 digits

SFF, UFF n n (to maximum of 31,then

truncated)

Y2C, Y2P, Y2S, Y2Z 2 4 digits

Table 27. Data Conversion Table

Syncsort MFX Programmer’s Guide 2–157

OUTREC

Data Conversion

Y2B, Y2D 1 4 digits

Y2ID 1 2 bytes

Y2IP 2 3 bytes

Table 27. Data Conversion Table

For full-date formats, the number of bytes in the input field can vary. The following
table shows input lengths for full-date formats and the resulting output length:

Input Format Bytes in Input Field Resulting Digits (d)

Y2T 3 5

4 6

5 7

6 8

Y2U 2 5

3 7

Y2V 3 6

4 8

Y2W 3 5

4 6

5 7

6 8

Y2X 2 5

3 7

Y2Y 3 6

4 8

Y4T 7 7

8 8

Y4W 7 7

8 8

Table 28. Data Conversion Table – Full-Date Formats

2–158 Syncsort MFX Programmer’s Guide

 OUTREC

Input Format Bytes in Input Field Resulting Digits (d)

Y4U 4 7

Y4V 5 8

Y4X 4 7

Y4Y 5 8

Table 28. Data Conversion Table – Full-Date Formats

For any other type of expression (those that are not a simple p,l,fi), MFX internally
maintains a 31-digit number. The number of digits that are used for conversion or
editing depends upon the lengths of the fields used in the expression. For more
information, see the description of expression in the section “The following
describes the c: subparameter:” on page 2-140.
If all fields in the expression conform to the following, then 15 digits will be used. If
any field in the expression exceeds these length values, 31 digits will be used for
editing or conversion. Note that full-date formats in an expression (not a simple
p,l,f) are treated as providing a 15-digit value when evaluating the following rules.

Fields in the expression Input Field Length

CSF/FS format fields 1 to 16 bytes

SFF/UFF format fields 1 to 15 bytes

BI/FI fields 1 to 4 bytes

PD fields 1 to 8 bytes

ZD fields 1 to 15 bytes

Decimal constants 1 to 15 significant digits

Table 29. Field Lengths That Produce a 15-Digit Default Output Length

If you specify no other FIELDS subparameters, the result will be converted to

printable output according to the default editing mask, M0. See “Mm
Subparameter (Editing Masks)” on page 2-179. Other forms of printable output can
be created by using the EDIT, LENGTH, Mm, and SIGNS subparameters, which
allow you to create your own edit patterns, or by using one of the 27 MFX-supplied
editing masks, which are appropriate for many editing operations.
To convert to a numeric data field, specify an output format of BI, CSF/FS, FD, FI,
PD, PDC, PDF, ZD, ZDC or ZDF. The default output field length is determined for

Syncsort MFX Programmer’s Guide 2–159

OUTREC

CSF/FS, PD, PDC, PDF, ZD, ZDC and ZDF formats by Table 30 on page 2-160 . For
BI and FI formats, use Table 31 on page 2-160. For FD, use Table 32 on page 2-161.
The number of digits (d) in the following table is obtained from column 3 of Table
27 on page 2-157 for an expression that is a single p,l,fi field. For any other type of
expression (not a single p,l,fi), d is either 15 or 31 based upon the fields in the
expression. If all fields in the expression conform to the lengths in Table 29 on page
2-159, then d is 15. If any are longer, d is 31.

Output Format Default Output Length (bytes)

CSF/FS d+1

PD, PDC, PDF d/2+1

ZD, ZDC, ZDF d

Table 30. Default Output Lengths

For BI or FI fields, when the field to be converted is a single p,l,fi field, the default
output length is either 4 or 8 bytes depending upon the format and length of the
field to be converted. See Table 31 on page 2-160 to determine the default length.

Input Format and Length Default Output Length (bytes)

BI or FI from 1 to 4 bytes 4

BI or FI from 5 to 8 bytes 8

CSF/FS from 1 to 16 bytes 4

CSF/FS from 17 to 32 bytes 8

FL either 4 or 8 bytes 8

PD from 1 to 8 bytes 4

PD from 9 to 16 bytes 8

SFF/UFF from 1 to 9 bytes 4

SFF/UFF from 10 to 44 bytes 8

Table 31. (Page 1 of 2)Output Lengths for BI and FI Formats when Input is a Single p,l,fi
Field

2–160 Syncsort MFX Programmer’s Guide

 OUTREC

Input Format and Length Default Output Length (bytes)

ZD from 1 to 15 bytes 4

ZD from 16 to 31 bytes 8

Table 31. (Page 2 of 2)Output Lengths for BI and FI Formats when Input is a Single p,l,fi
Field

These lengths can be overridden by specifying the LENGTH parameter.

For any other type of expression (those that are not simple p,l,fi), the output length
when converting to BI or FI is based upon the number of digits available for output.
If all fields in the expression conform to the lengths in Table 29 on page 2-159, then
d is 15. If any are longer, d is 31. When the number of digits available is 15, the
output length will be 4. When the number of digits is 31, the output length will be
8.
Table 32 on page 2-161 displays the formats that can be converted to FD and the
default output lengths.

Input Format and Length Default FD Format Output Length

(bytes)

BI or FI from 1 to 4 bytes 8

BI or FI from 5 to 8 bytes 16

CSF/FS from 1 to 16 bytes 8

CSF/FS from 17 to 32 bytes 16

FL either 4 or 8 bytes 16

PD from 1 to 8 bytes 8

PD from 9 to 16 bytes 16

SFF/UFF from 1 to 15 bytes 8

SFF/UFF from 16 to 44 bytes 16

ZD from 1 to 15 bytes 8

Table 32. (Page 1 of 2)Output Lengths for FD Fields

Syncsort MFX Programmer’s Guide 2–161

OUTREC

Input Format and Length Default FD Format Output Length

(bytes)

ZD from 16 to 31 bytes 16

Table 32. (Page 2 of 2)Output Lengths for FD Fields

If a LENGTH parameter is specified with a conversion to an FD format, only 4, 8

and 16 bytes are allowed.
The following five sections describe the data conversion capabilities:
• Converting Year Data with century window processing on INREC, OUTREC, or
OUTFIL OUTREC
• The EDIT Subparameter
• The LENGTH=n Subparameter
• The Mm Subparameter (Editing Masks)
• The SIGNS Subparameter

Converting Year Data with Century Window Processing on INREC,

OUTREC, or OUTFIL OUTREC
A 2-digit year-only field, as specified by the Y2B, Y2C, Y2D, Y2P, Y2S, Y2Z, Y2ID,
and Y2IP formats, can be converted on output to a 4-digit year.
The following describes output data conversion for 2-digit year-only date fields:
• The Y2B format specifies 2-digit, 1-byte binary year data that will be converted
to a 4-digit, displayable character format with the appropriate century value.
For information on the range of binary values representing year data with Y2B,
see Table 39 on page 2-239.
• The Y2C and Y2Z formats specify 2-digit year data that are in displayable
(zoned decimal) format. The 2-digit year data will be expanded to a 4-digit field
containing the appropriate century value.
• The Y2S format is equivalent to Y2C and Y2Z for valid numeric year data. All
three formats will convert such data to a displayable 4-digit year with the
appropriate century value. Y2S, however, provides additional functionality. For
data with binary zeros (X'00'), a blank (X'40') or binary ones (X'FF') in the first
byte, typically to identify header/trailer records, Y2S will expand the data to 4
bytes, padded in the first 2 bytes with the same character as found in the first
byte of the input field. The fourth byte of the output field is copied unchanged
from the second byte of the input field.
The following symbolic representation shows the treatment in hexadecimal of
the three types of data:

2–162 Syncsort MFX Programmer’s Guide

 OUTREC

SORTIN Input OUTREC Output

00ab 000000ab

40ab 404040ab

FFab FFFFFFab

• The Y2D and Y2P formats specify 2-digit year values in packed decimal format.
The processing applied to these fields will create a 4-digit year value converted
to a displayable character format.
• The Y2ID and Y2IP formats take as input the same 2-digit packed decimal year
data as the Y2D and Y2P formats but produce a 4-digit year output that
remains in packed decimal format. Y2ID will convert data from X'yy' to X'ccyy',
and Y2IP will convert data from X'ayys' to X'accyys', where cc is the correct
century. (For a description of Y2D and Y2P formats, see “The Y2D Format” on
page 2-240 and “The Y2P Format” on page 2-240.
For full-date fields with 2-digit years (Y2T, Y2U, Y2V, Y2W, Y2X, and Y2Y), the 2-
digit portion will expand to the appropriate 4-digit year based on the CENTWIN
setting. When doing a simple p,l,f conversion to a printable format, the output field
length can be determined from Table 28 on page 2-158. Conversion to other formats
can be done using the DT, DTNS, TOJUL or TOGREG parameters.
Note that an additional data format, PD0, which is typically used to process the
month and day portion of packed decimal data, is not affected by CENTWIN
processing and will not convert 2-digit year data to 4-digit years. PD0 can be used
with the MFX-supplied edit mask M11. The year data formats Y2B, Y2C, Y2D, Y2P,
and Y2Z or the full-date formats Y2T, Y2U, Y2V, Y2W, Y2X, and Y2Y can also be
used when forming expressions. The 4-digit year for year data formats or the full
date data for the full-date formats will be converted to an integer for arithmetic
calculations. Any expression with these formats can also be converted to an output
numerical data format fo or to printable output by specifying one or more of the
OUTREC FIELDS subparameters (Mm, EDIT, SIGNS, or LENGTH). For
information on using the year data formats for SORT or MERGE field
specifications, see “CENTWIN Parameter (Optional)” on page 2-239 or “CENTWIN
Parameter (Optional)” on page 2-71, respectively. For more information on using
the year data formats for INREC or OUTREC processing, see “Example 5” on
page 2-222.
For more information on converting full-date formats, see the descriptions of the fi
and fy2f,(c) parameters on pages 2-142-2-147, Table 20 on page 2-148, and Table 28
on page 2-158.

Syncsort MFX Programmer’s Guide 2–163

OUTREC

Converting SMF Date and Time Formats

You can convert SMF date and time formats to standard date and time formats.
The following table shows the SMF formats and the converted output:

SMF Format Converted Output

DT1 Z'yyyymmdd'

DT2 Z'yyyymm'

DT3 Z'yyyyddd'

TM1 Z'hhmmss'

TM2 Z'hhmm'

TM3 Z'hh'

TM4 Z'hhmmssxx'

Table 33. SMF Formats and Converted Output

For DTn, the source is the 4-byte packed SMF date value (P'cyyddd'). For TMn, the
source is a 4-byte binary SMF time value.
The c in the date source P'cyyddd' represents the century. It is converted as follows:
0 is converted to 19, 1 is converted to 20, and 2 or greater is converted to 21.
The converted output is a zoned decimal field, where each character in the table
represents a single byte. For TM4, xx represents hundredths of a second.
The MFX predefined edit masks (M0-M26) or specified edit patterns can be used to
edit the converted date and time. The default mask is M11.
Note: A data exception (0C7 ABEND) or an inaccurate ZD date can occur if an
SMF date is not valid. An inaccurate ZD time can occur if an SMF time is not valid.
SMF dates and times are processed as positive values.
For an example of an OUTREC control statement that converts SMF formats, see
Figure 123 on page 2-222.

Time of Day Formats (DCn, TCn, DEn, and TEn)

The time of day data (TOD) and time of day data extended format (ETOD) created
by the STCK or STCKE hardware instruction can be interpreted to produce several
variations of date and time values. The formats DCn, TCn, DEn, and TEn specify
what information is to be extracted from the TOD value and presented for use in an
expression. This data can either be directly used for conversion to a printable value
or another data format or used as an individual term in an expression. The system
service STCKCONV is used to perform the conversion to the desired format.
The following table describes the available formats and the output format of the
data that is extracted from the TOD or ETOD field. In all cases, the input is an 8-

2–164 Syncsort MFX Programmer’s Guide

 OUTREC

byte field. For TOD formats (DCn and TCn), the entire TOD field is used. For
ETOD formats (DEn and TEn), the position specified should reflect the first 8 bytes
of the ETOD field.

Format Converted Output

DC1 Z’yyyymmdd’

DC2 Z’yyyymm’

DC3 Z’yyyyddd’

DE1 Z’yyyymmdd’

DE2 Z’yyyymm’

DE3 Z’yyyyddd’

TC1 Z’hhmmss’

TC2 Z’hhmm’

TC3 Z’hh’

TC4 Z’hhmmssxx’

TE1 Z’hhmmss’

TE2 Z’hhmm’

TE3 Z’hh’

TE4 Z’hhmmssxx’

Table 34. Time of Day Formats and Converted Output

yyyy represents a four digit year.

mm represents the month (01-12).
dd represents the day of the month (01-31).
ddd represents the day of the year (001-366).
hh represents the hour (00-23).
mm represents the minutes (00-59).
ss represents seconds (00-59).
xx represents hundredths of a second (00-99).

Syncsort MFX Programmer’s Guide 2–165

OUTREC

The MFX predefined edit masks (M0-M26) or specified edit patterns can be used to
edit the converted date and time. The default mask is M11.

Literal Fields Subparameters

Spaces (X), hexadecimal digits (X'hhhh...hh'), literal strings (C'literal string'), and
binary zeros (Z) can also be specified in the FIELDS parameter. Each of these
entries can be preceded by an 'n' value which indicates that a specified number of
spaces, hex digits, literal strings, or binary zeros should be inserted in the output
record. Additionally, you can insert the date and time, or the date with an offset, of
your MFX run into your records.
nX Use the nX entry to specify a number n of spaces. The n value
may be any number from 1 to 4095 inclusive. The X entry rep-
resents a space and must be coded to the immediate right of
the number specified for n. If more than 4095 spaces are
desired, two or more nX values should be specified.
nX'hhhh...hh' Use the nX'hhhh...hh' entry to specify that n copies of hex dig-
its or hex digit strings should be inserted in the output record.
(Each hh pair is 1 byte of output.) The repetition factor n may
be any number from 1 to 4095 inclusive.
nC'literal string' Use the nC'literal string' entry to specify that n copies of lit-
eral strings should be inserted in the output record. The repe-
tition factor n may be any number between 1 and 4095
inclusive. An apostrophe within a literal string must be speci-
fied with a double apostrophe (e.g., C'O''LEARY').
nZ Use the nZ entry to define a specified number n of binary
zeros that will be inserted in the output record. The repetition
factor n may be any number between 1 and 4095 inclusive.
The Z entry must be coded to the immediate right of n.

Generating Run-time Date and Time Constants

You can insert the date and time, or the date with an offset, of your MFX run into
your records. Table 35 (“Run-time Constants”) on page 2-167 shows the constants
generated by the run-time date and time parameters.
A 'C' in the output format denotes a character constant. A 'P' denotes a packed
decimal constant, which contains a positive sign and a leading zero when padding
is necessary. A '(c)' in the parameter represents a separator character.
Optionally, you can create an offset of the current date. The offset takes the form
{±}nnnn, where '+' indicates a date after the current date and '–' indicates a date
before the current date. 'nnnn' is the date offset. The range is 0 to 9999, which
represents the number of days to be added or subtracted from the current date; or 0

2–166 Syncsort MFX Programmer’s Guide

 OUTREC

to 999, which represents the number of months to be added or subtracted from the
current month for DATE2, DATE2P, or DATE2(c).

Parameter Output Length (Bytes)

&DATE [{±}nnnn] C'mm/dd/yy' 8

&DATE1 [{±}nnnn] C'yyyymmdd' 8

&DATE1(c) [{±}nnnn] C'yyyycmmcdd' 10

&DATE1P [{±}nnnn] P'yyyymmdd' 5

&DATE2 [{±}nnn] C'yyyymm' 6

&DATE2(c) [{±}nnn] C'yyyycmm' 7

&DATE2P [{±}nnn] P'yyyymm' 4

&DATE3 [{±}nnnn] C'yyyyddd' 7

&DATE3(c) [{±}nnnn] C'yyyycddd' 8

&DATE3P [{±}nnnn] P'yyyyddd' 4

&DATE4 [{±}nnnn] C'yyyy-mm-dd-hh.mm.ss' 19

&DATE5 [{±}nnnn] C'yyyy-mm-dd-hh.mm.ss.nnnnnn' 26

&DATE=(m1m2m3m4) [{±}nnnn] (see description below table)

&DATENS=(xyz) [{±}nnnn] (see description below table)

&TIME C'hh:mm:ss' 8

&TIME1 C'hhmmss' 6

&TIME1(c) C'hhcmmcss' 8

&TIME1P P'hhmmss' 4

&TIME2 C'hhmm' 4

&TIME2(c) C'hhcmm' 5

&TIME2P P'hhmm' 3

&TIME3 C'hh' 2

&TIME3P P'hh' 2

&TIME=(hp) (see description below table)

&TIMENS=(tt) (see description below table)

&YDDD=(m1m2m3) [{±}nnnn] (see description below table)

Table 35. Run-time Constants

Syncsort MFX Programmer’s Guide 2–167

OUTREC

Parameter Output Length (Bytes)

&YDDDNS=(m1m2) [{±}nnnn] (see description below table)

Table 35. Run-time Constants

&DATE=(m1m2m3m4)[{±}nnnn] This form of the &DATE subparameter gener-

ates the current system date or date with offset and controls
the formatting of the date. You can specify the position of
the year, month, and day; specify a separator character; and
choose between 2-digit and 4-digit year representation.
The positions m1 through m4 represent masks used to for-
mat the date. To specify the positions of the month, day, and
year, replace the m1, m2 and m3 positions, in any order, with
M for the month (01-12), D for the day (01-31), and either Y
or 4 for the year (where Y is a 2-digit year and 4 is a 4-digit
year). Replace the m4 position with a separator character.
For example, to print the date with the form yy-mm-dd,
specify &DATE=(YMD-). For December 31, 1997, the date
would appear as “97-12-31.”
The field for this form of &DATE requires 8 bytes for a 2-
digit year representation and 10 bytes for a 4-digit year. The
M, D, and Y or 4 may only appear once in the mask. All four
positions must be specified.
Optionally, you can create an offset of the current date. See
“Generating Run-time Date and Time Constants” on page 2-
166 for a description.
&DATENS=(xyz)[{±}nnnn] This form of the &DATENS subparameter specifies
that the current date or date with offset is to appear in the
output record in the form 'xyz', where x, y, and z indicate the
order in which the month, day, and year are to appear and
whether the year is to appear as two or four digits. For x, y,
and z, use M to represent the month (01-12), D to represent
the day (01-31), Y to represent the last two digits of the year
(for example, 02), or 4 to represent the four digits of the year
(for example, 2002). M, D, and Y or 4 can each be specified
only once.
For example, &DATENS=(DMY) would produce a date of
the form 'ddmmyy' which on March 29, 2002, would appear
as '290302'. &DATENS=(4MD) would produce a date of the
form 'yyyymmdd' which on March 29, 2002, would appear as
'20020329'. x, y, and z must be specified.

2–168 Syncsort MFX Programmer’s Guide

 OUTREC

Optionally, you can create an offset of the current date. See

“Generating Run-time Date and Time Constants” on page 2-
166 for a description.
&TIME=(hp) This form of the &TIME subparameter generates the cur-
rent system time of day and controls the formatting of the
time. You can print the time in 24-hour or 12-hour formats
and specify the separator character between the hours, min-
utes and seconds.
The format for 24-hour time is hhpmmpss, where hh rep-
resents the hour (00-23), mm represents minutes (00-59), ss
represents seconds (00-59), and p represents the separator
character as specified by p in the &TIME=(hp) subparame-
ter.
The format for 12-hour time is hhpmmpss nn, where hh rep-
resents the hour (01-12), mm represents minutes (00-59), ss
represents seconds (00-59), and p represents the separator
character as specified by p in the &TIME=(hp) subparame-
ter. The nn is “am” or “pm” as appropriate.
To select 12-hour mode specify h as 12; to select 24-hour
mode specify h as 24. The p specification represents the
character to use as a separator.
For example, to display the time in a 12-hour format with a
period as a separator, specify &TIME=(12.). At 22:43:23
hours, the time would appear as “10.43.23 pm.”
The field for this form of the &TIME subparameter requires
8 bytes for the 24-hour format and 11 bytes for the 12-hour
format.
&TIMENS=(tt) This form of the &TIMENS subparameter specifies that the
current time is to appear in the output record in the form
'hhmmss' (24-hour time) or 'hhmmss xx' (12-hour time). If tt
is 24, the time is to appear in the form 'hhmmss' (24-hour
time) where hh represents the hour (00-23), mm represents
the minutes (00-59), and ss represents the seconds (00-59).
For example, &TIMENS=(24) would produce a time of the
form 'hhmmss' which at 08:25:13 pm would appear as
'202513'. If tt is 12, the time is to appear in the form
'hhmmss xx' (12-hour time) where hh represents the hour
(01-12), mm represents the minutes (00-59), ss represents
the seconds (00-59), and xx is either 'am' or 'pm'.
For a second example, &TIMENS=(12) would produce a
time of the form 'hhmmss xx' which at 08:25:13 pm would
appear as '082513 pm'.

Syncsort MFX Programmer’s Guide 2–169

OUTREC

&YDDD=(m1m2m3)[{±}nnnn] This form of the &YDDD subparameter specifies

that the current date or date with offset is to appear in the
output record in the form of a year and day. You can specify
the position of the year and day, specify a separator charac-
ter, and choose between 2-digit and 4-digit year representa-
tion. The positions m1 through m3 represent masks used to
format the date. To specify the position of the year and day,
replace the m1 and m2 positions (in either position) with D
for day (001-366) and either Y or 4 for the year (where Y is a
2-digit year and 4 is a 4-digit year). Replace the m3 position
with a separator character.
For example, to print the date in the form yyyy/ddd, specify
&YDDD=(4D/). For March 29, 2005, the date would appear
as 2005/088.
Optionally, you can create an offset of the current date. See
“Generating Run-time Date and Time Constants” on page 2-
166 for a description.
&YDDDNS=(m1m2)[{±}nnnn] This form of the &YDDDNS subparameter speci-
fies that the current date or date with offset is to appear in
the output record in the form of a year and day. You can
specify the position of the year and day and choose between
2-digit and 4-digit year representation. The positions m1
and m2 represent masks used to format the date. To specify
the position of the year and day, replace the m1 and m2 posi-
tions (in either position) with D for day (001-366) and either
Y or 4 for the year (where Y is a 2-digit year and 4 is a 4-
digit year).
For example, to print the date in the form dddyy, specify
&YDDDNS=(DY). For March 29, 2005, the date would
appear as 08805.
For an example of an OUTREC control statement that gen-
erates run-time constants, see Figure 124 on page 2-223.
Optionally, you can create an offset of the current date. See
“Generating Run-time Date and Time Constants” on page 2-
166 for a description.

Function Field Subparameters

The function fields allow you to insert a sequence number into your output record,
to add or subtract values from date fields, to compute the difference between two
date fields, and perform other date field functions.
The figure below illustrates how the function subparameters should be specified
and describes their function,

2–170 Syncsort MFX Programmer’s Guide

 OUTREC

 1  1   (p,h)  
 SEQNUM,1,f ,START=  ---  ,INCR=  ---  ,RESTART=   
 n
  i
   ( %pp )  
 
 DATEADD=(datefield,number,unit) 
 DATEDIFF=(datefield 1 ,datefield 2 ,unit) 
 
  ADDDAYS  
  ADDMONS  
  ADDYEARS   TOGREG= f o [ ( c ) ]  
 p,l,f yxx ,  
[c:]   SUBDAYS 
,numeric_field, 
 TOJUL= f o [ ( c ) ] 


  SUBMONS  
  SUBYEARS  
 
  NEXTDday  
  PREVDday  
  LASTDAYW   TOGREG= f o [ ( c ) ]  
 p,l,f yxx,  LASTDAYM  ,  
    TOJUL= f o [ ( c ) ]  
  LASTDAYQ  
  LASTDAYY  
 

Figure 75. Function Field Subparameters

SEQNUM Use SEQNUM to create a sequence number field within the

output record. The length of the field can be from 1 to 16 bytes
and can be represented in either BI, PD, or ZD formats. A
starting value and an increment can be specified for the field.
In addition, the sequence numbering can be restarted when
the value in a specified field changes.
The following describes the SEQNUM variables and parame-
ters:
l Represents the length in bytes of the field to be
created. A value from 1 to 16 can be specified.
f Indicates the format of the field to be created.
BI, PD, or ZD can be specified to create an
unsigned binary field, a packed decimal field, or
zoned decimal field, respectively.
START Optionally specifies a starting number n for the
field. The n value can be 0 through
2,147,483,647. The default is 1.
INCR Optionally specifies a value i that indicates
how sequence numbers should be incremented.
The i value can be 1 through 65,535. The
default is 1.
RESTART Optionally specifies that the sequence number-
ing is restarted when the value in the
RESTART field changes. The value of n speci-
fied in the START parameter is used to restart

Syncsort MFX Programmer’s Guide 2–171

OUTREC

the numbering sequence. p represents the posi-

tion of the first byte of the field. h is the length
of the field and can be from 1 to 256 bytes.
Alternatively, a parsed field %pp may be speci-
fied. A binary comparison is performed on the
field.
The maximum sequence number generated is limited to 15
decimal digits or the output field length. If a number is
reached that would exceed the limit, MFX truncates the high-
order digit and continues processing. Thus, sequence numbers
will cycle within the limit. So, if the output is a 2-byte ZD
field, 99 will be the highest sequence number. The next num-
ber, 100, will have its high-order digit truncated. The result-
ing number, 00, starts a new sequence number cycle from 00
to 99, regardless of the START value.
DATEADD Use DATEADD to add or subtract units of days to or from an
input record date field and create an output record date field
in the same format with the same length. The results of
DATEADD are considered to be a character string and not a
number. Thus they cannot be used where a numeric field
could be used, for example, in an arithmetic expression or
with an EDIT pattern or mask.
The following describes the DATEADD parameters:
datefield Defines a date field in the input record. The for-
mat may be printable or packed, with or with-
out a separator, containing either a 2-digit or 4-
digit year, and either a Julian day, quarter, or
month/day. If a 2-digit year is specified,
CENTWIN processing will be used to deter-
mine a 4-digit year, but the final result will only
contain a 2-digit year. The field is defined by
specifying a position and length followed by the
DT, DTNS, or DTNSP parameter. The field
length must be exactly what is implied by the
mask specifications.
When there is invalid data, the output field will
contain Z’9999...’ for a printable field or
X’9999...C’ for a packed field.
p,l,DT[=({m1m2m3m4,m1m2m4})] DT is used for
printable fields with separators. Position
and length of the input field are followed by
the DT parameter which describes the input
format. For day/month/year fields, m1

2–172 Syncsort MFX Programmer’s Guide

 OUTREC

through m3 are replaced, in any order, with

M for the month (01-12), D for the day (01-
31), and either Y or 4 for the year (where Y
is a 2-digit year and 4 is a 4-digit year); m4
designates the separator character. For
quarter/year, month/year and Julian
day/year, just specify m1m2 in any order,
with Y or 4 for the year and either Q for the
quarter, J for the 3-digit Julian day or M for
the month; m4 designates the separator
character. If the mask specification is omit-
ted, a mask of ‘MDY/ ’ will be assumed.
p,l,DTNS[=({m1m2m3,m1m2})] DTNS is similar to
DT, but is used for printable fields without
separators. If the mask specification is
omitted, a mask of ‘MDY’ will be assumed.
p,l,DTNSP[=({m1m2m3,m1m2})] DTNSP is similar
to DTNS, but is used for packed decimal
fields without separators. Fields are
assumed to contain a trailing sign which is
ignored. If the mask specifies an even num-
ber of digits for the date, then the input
field is assumed to contain a leading half-
byte of zeros which is ignored.
number Specifies the number of date units to be added
to or subtracted from the input date. Specify a
positive number to add or a negative number to
subtract. The absolute value of the number is
limited to 9999 for DAYs and WEEKs and 999
for MONTHs, QUARTERs and YEARs.
unit Specifies the date unit for the calculation. If the
mask includes a day/month, then specify DAY,
WEEK or YEAR as the unit. If the mask
includes a Julian day, then specify DAY or
WEEK as the unit. If the mask includes a quar-
ter or month without a day, then specify
MONTH, QUARTER or YEAR as the unit.

Syncsort MFX Programmer’s Guide 2–173

OUTREC

Sample DATEADD Specifications

Example 1

OUTREC FIELDS=(.....DATEADD=(10,8,DT=(J4-),13,WEEK),.....)

Figure 76. Sample DATEADD Subparameter

Column 10 contains a Julian date in the form ddd-yyyy and 13 weeks are to be
added to it to create the output field.
Example 2

OUTREC FIELDS=(.....DATEADD=(10,4,DTNSP=(YMD),-30,DAY),.....)

Figure 77. Sample DATEADD Subparameter

Column 10 contains a Gregorian date in the form X’0yymmdds’ and 30 days are to
be subtracted from it to create the output field.
Example 3

OUTREC FIELDS=(.....DATEADD=(10,6,DTNS=(M4),3,MONTH),.....)

Figure 78. Sample DATEADD Subparameter

Column 10 contains a month/year date field in the form mmyyyy and 3 months are
to be added to it to create the output field.
DATEDIFF Use DATEDIFF to compute the interval between two date
values. It returns a numeric value that can be used in expres-
sions or formatted with an EDIT pattern; the default format-
ting will be with mask M0 for 15 digits, generating a 16-byte
field. The value is calculated by counting the number of unit
boundaries between the two date values. Therefore, there is
no rounding. For example, when unit is DAY, the value is the
number of midnights between datefield1 and datefield2 - mid-
night is the boundary between one day and the next. Simi-
larly, the week boundary is defined as midnight on Sunday;
the month boundary is midnight of the last day of the month,
and so on.
The following describes the DATEDIFF parameters:
datefield1 Contain the two date values to be used in the
calculation.
datefield2 Date values may be date fields within the
record, a hard-coded date, or the current sys-
tem date, though each of the two fields need not

2–174 Syncsort MFX Programmer’s Guide

 OUTREC

be in the same format. Date fields within the

record are specified in the same manner as the
datefield in the DATEADD function, hardcoded
dates must be of the form YYYY/MM/DD
(length=10), and the current system date is
indicated by &DATE.
If datefield1 is earlier than datefield2, the
value is negative; otherwise, the value is posi-
tive. When there is invalid datefield data in
either field, the value will be
Z’999999999999999’, and an informational
WER490I message will be issued.
unit Specifies the date unit to be used in calculating
the difference between datefield1 and date-
field2. Specify YEAR, QUARTER, MONTH,
WEEK, or DAY. The date components required
for the calculations should be present in both
datefields. For MONTH, the month is required
in each datefield. For WEEK and DAY, the day
is required.

Sample DATEDIFF Specifications

Example 1

OUTREC FIELDS=(.....DATEDIFF=(10,8,DT=(J4-),20,4,DTNSP=(YDM),

DAY),.....)

Figure 79. Sample DATEDIFF Subparameter

Column 10 contains a Julian date in the form ddd-yyyy, column 20 contains a

Gregorian date in the form X’0yyddmms’ and the difference in days is desired.
Example 2

OUTREC FIELDS=(.....DATEDIFF=(10,6,DTNS=(YMD),20,5,DT=(YM/),

MONTH),M12,LENGTH=5,.....)

Figure 80. Sample DATEDIFF Subparameter

Column 10 contains a Gregorian date in the form yymmdd and column 20 contains
a year/month date in the form yy/mm. The difference in months is desired,
although it is not known which the earlier date is (a negative result would indicate
that it was datefield1). The result will be in a 5-byte field with a leading sign, using
the M12 editing mask.

Syncsort MFX Programmer’s Guide 2–175

OUTREC

Example 3

INREC FIELDS=(.....DATEDIFF=(2013/06/08,&DATE,DAY),.....)

Figure 81. Sample DATEDIFF Subparameter

The number of days from the current date until June 8th, 2013 is needed.
For the following date arithmetic functions, the eligible p,l,f fields are the full-date
fields from Table 28 on page 2-158.

p1,l1,f1yxx,DATEDIFF,p2,l2,f2yxx

Figure 82. Alternate DATEDIFF usage for full-date fields

This is an alternate usage of DATEDIFF for full-date fields only. This alternate
usage produces the difference between the two dates in days only, and the output is
a printable field rather than a numerical value. The output field is 8 bytes, where
the first byte is a + or - sign byte, followed by 7 digits. The sign byte will be + if the
first field is greater than or equal to the second field, otherwise the byte will be - . A
parsed field can be used for p1,l1 but not for p2,l2. .

 ADDDAYS 
 ADDMONS 
 ADDYEARS   TOGREG= f o [ ( c ) ] 
p,l,f yxx,  ,numeric_field,  
 SUBDAYS   TOJUL= f o [ ( c ) ] 
 SUBMONS 
 SUBYEARS 

Figure 83. Date Functions

These date functions are used to add or subtract any of days, months, or years from
the input field to create a formatted output date field.
When the unit is months or years, a valid output date will always be created. For
example, subtracting one month from a March 30th date will result in February 28th
in a non-leap year and February 29th in a leap year.
The numeric_field can be a number (+n or -n) or a numeric field (p,l,f) in the record,
where f can be BI, FI, ZD, PD, FS, UFF, or SFF with a valid length for those
formats. The value of the numeric_field must be from -3652058 to +3652058 for
ADDDAYS or SUBDAYS, from -119987 to +119987 for ADDMONS or SUBMONS,
or from -9998 to +9998 for ADDYEARS or SUBYEARS.

2–176 Syncsort MFX Programmer’s Guide

 OUTREC

The output field is formatted by either the TOJUL or TOGREG parameter as

described on page 2-154.

 
 NEXTDday 
 PREVDday 
 LASTDAYW   TOGREG= f o [ ( c ) ] 
p,l,f yxx,  LASTDAYM  , 
   TOJUL= f o [ ( c ) ] 
 LASTDAYQ 
 LASTDAYY 
 

Figure 84. Date Functions

These date functions are used to compute new calendar dates based on a full-date
input field. The input p,l can also be a parsed field.
The NEXTDday function computes the next specified weekday date relative to the
input date. ''day'' in the keyword can be any of SUN, MON, TUE, WED, THU, FRI
or SAT, such as NEXTDSAT.
The PREVDday function computes the previous specified weekday date relative to
the input date. See above for valid ''day'' values.
LASTDAYW computes the Friday date of the week of the input date, where Friday
is considered the last day of the week.
LASTDAYM computes the last date of the month of the input date.
LASTDAYQ computes the last date of the quarter of the input date.
LASTDAYY computes the last date of the year of the input date.
The output field is formatted by either the TOJUL or TOGREG parameter as
described on page 2-154.

EDIT Subparameter
The EDIT subparameter lets you create your own edit patterns for converted
numeric data. An edit pattern can consist of:
• Significant digit selectors.
• Leading insignificant digit selectors.
• Sign replacement characters.
• Any other characters to be printed in the actual output.
The edit pattern can be up to 44 characters in length, with a maximum of 31 digits.
The characters used to represent significant or insignificant digit selectors are
determined by the keyword EDIT. If EDIT is specified, the letter I represents
leading insignificant digits which will print as blanks if the digits are zeros, and
the letter T represents significant digits (digits that will print in their true form,
even as leading zeros).

Syncsort MFX Programmer’s Guide 2–177

OUTREC

The keyword EDIT can be specified with replacements for the letters I and/or T.
Any printable character can be used as a replacement character. This replacement
provides you with a pattern that encompasses all printable characters.
The figure below illustrates the concept of replacing the insignificant and
significant digit selectors I and T with other characters.

EDxy=

where:

x = insignificant digit selector

y = significant digit selector

Figure 85. Replacing Digit Selector Characters

When a blank, quotation mark or unbalanced parenthesis appears within an EDIT

pattern, the entire pattern must be enclosed within single quotation marks.
Balanced parentheses need not be enclosed within quotation marks. A single
quotation mark within the pattern (i.e., an apostrophe) must be specified as two
apostrophes.
All other characters are printed as specified in the edit pattern, with the following
exceptions:
• Any character specified after the first leading insignificant digit selector and
before the first significant digit selector will print as a blank, unless a
previously selected digit was non-zero.
• Any character specified after the last significant digit selector will print as a
blank if the edited number is positive.
• Any character or character string specified before the first leading insignificant
digit selector, including a leading sign character, will print to the immediate left
of the first significant digit. The appropriate number of leading blanks will be
supplied, assuring that the total number of characters in the printed field
corresponds to the total number of characters in the edit pattern.
• Any leading insignificant digit selector specified after the first significant digit
selector will be treated as a significant digit selector.
• The sign replacement character appearing as the first and/or last character of
the pattern is replaced as per the SIGNS subparameter.

LENGTH=n Subparameter
Use the LENGTH=n subparameter to alter the default length of the output field
data. The maximum value which can be specified for n is 44.
• When an editing mask is used, the default length is determined by the edit
pattern and the format of the field. If LENGTH=n is not specified, the length is

2–178 Syncsort MFX Programmer’s Guide

 OUTREC

equal to the number of characters specified in the edit pattern. If LENGTH=n is

specified, the edit pattern will either be truncated on the left or padded with
blanks on the left so that the length of the pattern equals the n value.
The maximum value that can be specified for n when editing masks are used is
44.
• When output data format fo is used, the default length is determined based on
the expression characteristics. For more information, see “How to Convert
Numeric Data” on page 2-157. If LENGTH=n is specified, the output data will
either be truncated on the left or padded on the left with zeros (or blanks for
CSF/FS) of the appropriate format to a length of n. For FD output format, only
4, 8 or 16 may be specified for the length.
For other output formats, the maximum value that can be specified for n when
an output data format fo is used is 44.

Mm Subparameter (Editing Masks)

MFX provides editing masks to simplify the more common editing operations.

Syncsort MFX Programmer’s Guide 2–179

OUTREC

Mask Pattern Signs Length

M0 IIIIIIIIIIIIIITS (,,' ',-) d+1

M1 TTTTTTTTTTTTTTTS (,,' ',-) d+1

M2 I,III,III,III,IIT.TTS (,,' ',-) d+1 + [d/3]

M3 I,III,III,III,IIT.TTCR d+2 + [d/3]

M4 SI,III,III,III,IIT.TT (+,-) d+1 + [d/3]

M5 SI,III,III,III,IIT.TTS (' ',(,' ',)) d+2 + [d/3]

M6 III-TTT-TTTT 12

M7 TTT-TT-TTTT 11

M8 IT:TT:TT 8

M9 IT/TT/TT 8

M10 IIIIIIIIIIIIIIT d

M11 TTTTTTTTTTTTTTT d

M12 SIII,III,III,III,IIT (' ',-) d+1 + [(d-1)/3]

M13 SIII.III.III.III.IIT (' ',-) d+1 + [(d-1)/3]

M14 SIII III III III IITS (' ',(,' ',)) d+2 + [(d-1)/3]

M15 III III III III IITS (,,' ',-) d+1 + [(d-1)/3]

M16 SIII III III III IIT (' ',-) d+1 + [(d-1)/3]

M17 SIII'III'III'III'IIT (' ',-) d+1 + [(d-1)/3]

M18 SI,III,III,III,IIT.TT (' ',-) d+1 + [d/3]

M19 SI.III.III.III.IIT,TT (' ',-) d+1 + [d/3]

M20 SI III III III IIT,TTS (' ',(,' ',)) d+2 + [d/3]

M21 I III III III IIT,TTS (,,' ',-) d+1 + [d/3]

M22 SI III III III IIT,TT (' ',-) d+1 + [d/3]

M23 SI'III'III'III'IIT.TT (' ',-) d+1 + [d/3]

M24 SI'III'III'III'IIT,TT (' ',-) d+1 + [d/3]

M25 SIIIIIIIIIIIIIIT (' ',-) d+1

M26 STTTTTTTTTTTTTTT (+,-) d+1

Table 36. Editing Masks

2–180 Syncsort MFX Programmer’s Guide

 OUTREC

Notes:
• If neither Mm nor EDIT is specified, M0 is used to edit BI, FI, FL, PD, PD0, ZD,
and CSF/FS fields and M11 is used to edit DTn, DCn, DEn, TMn, TCn, and TEn
fields.
• The letter d represents the number of resulting digits after data conversion.
The mask patterns in the Pattern column shows the resulting digits when the
number of digits is 15. (See Table 27 on page 2-157.) When the number of digits
to be displayed is greater than 15, the masks will be extended on the left with
the required digit selectors and constant characters.
• The bracket symbols indicate that only the integer part of this division should
be retained.
Table 36 on page 2-180 illustrates the following for each of the available masks.
• Edit pattern.
• Leading or trailing signs, where appropriate.
• Length. If an MFX editing mask is used for totaled or subtotaled data, the
length of the output field is determined from the length of the field and by using
Table 18 on page 2-112, not by the specified length of the input field. The
subparameter LENGTH can be used to override the length of the output field.
The edit patterns use the same symbolic letters used in the EDIT subparameter.
Leading insignificant digits are represented by the letter I; significant digits are
represented by the letter T. Leading or trailing sign replacement characters are
represented by the letter S. All other characters print as they appear in the
pattern.
The SIGNS illustrated for each mask follow the format requirements of the SIGNS
subparameter. You can specify the SIGNS subparameter to selectively override the
signs for a particular mask. For example, if you specify mask M4 and also specify
SIGNS=(' '), a leading blank will print instead of a plus sign if the number is
positive. However, a leading minus sign will print if the number is negative because
the leading negative sign specified in the editing mask has not been overridden.
The lengths in the table represent the length, in bytes, of the mask. The lengths of
masks M0-M5 and M10-M26 are determined, in part, by the number of digits d.
See Table 27 on page 2-157 to determine the number of digits for each type of
numeric field.

SIGNS Subparameter
The SIGNS subparameter specifies the sign(s) that will appear before or after the
edited number.
The sign replacement character, normally 'S', has special meaning if it appears as
the first or last character in an edit pattern. In these positions, the sign
replacement character will be replaced, as appropriate, by the characters specified
by the SIGNS subparameter.

Syncsort MFX Programmer’s Guide 2–181

OUTREC

The format of the SIGNS subparameter is illustrated below.

SIGNS=(s1,s2,s3,s4)

Figure 86. SIGNS Format

where:
s1= leading positive sign indicator
s2= leading negative sign indicator
s3= trailing positive sign indicator
s4= trailing negative sign indicator
Because the SIGNS subparameter contains four positional values, commas must be
used to indicate embedded, unspecified values. Each of the four values can contain
one, and only one, character; specified characters must be separated by commas.
A blank, comma, quotation mark and unbalanced parenthesis used as a SIGNS
character must be enclosed within apostrophes. An apostrophe used as a SIGNS
character must be specified as two apostrophes enclosed within apostrophes ('''').
When the SIGNS subparameter is specified, the letter 'S' is normally used as the
sign replacement character in the user-supplied edit pattern. You can change the
last letter of the keyword SIGNS in order to specify another character as the sign
replacement character. For example, if you specify SIGNX instead of SIGNS, the
letter 'X' becomes the sign replacement character in the user-provided edit pattern.
If you specify a sign replacement character in the edit pattern but do not specify a
value in the corresponding position in the SIGNS parameter, a blank will be
assumed. For example, if you specify the following:

EDIT=(IITT.TTS),SIGNS=(,,,-)

Figure 87. Sample EDIT Statement

then a trailing minus sign will print if the number is negative and a trailing blank
will print if the number is positive.
The SIGNS subparameter can also be used to override the sign values in MFX-
provided editing masks.

CHANGE Subparameter
The CHANGE subparameter changes an input field to a replacement value in the
reformatted output record if a specified field equals a search constant.
The format of the CHANGE subparameter is shown below:

2–182 Syncsort MFX Programmer’s Guide

 OUTREC

 p,l 
[c:]  ,CHANGE=(o,srch 1 ,repl 1 [,srch 2 ,repl 2 ,...srch n, repl n ] )
 %pp 
 nmrepl 
,NOMATCH=(  r,n )
 %pp 

Figure 88. CHANGE Subparameter

Multiple search-replacement paired values, with different data formats, can be
specified on a CHANGE subparameter. Note the following rules for mixing data
formats:
• Search constants are character, hexadecimal, or binary strings. Multiple search
constants on a CHANGE subparameter can be a mixture of character and
hexadecimal formats. Binary search constants cannot be mixed with search
constants of other formats; thus, if one search constant on a CHANGE
subparameter is binary, all other search constants on that subparameter must
also be binary.
• Replacement values are either character or hexadecimal string constants or a
field from the input record. Multiple replacement constants on a CHANGE
subparameter can be a mixture of character and hexadecimal string constants
and fields from the input record.
• The constants of a search-replacement pair can be of different data format. For
example, a hexadecimal or binary search constant could be paired with a
character replacement constant, or a character search constant could be paired
with a hexadecimal replacement constant. Thus, you could change a
hexadecimal or binary input field to a character output field, or you could
change a character input field to a hexadecimal output field.
The following describes the elements of the CHANGE subparameter:
p,l The normal MFX position-length designation that specifies the
search field. When this search field matches a search constant, the
input field will be changed in the output to a replacement value.
For character or hexadecimal search constants, the search field can
be 1 to 64 bytes long. For binary search constants, the search field
must be one byte.
%pp Identifies a fixed-length parsed field that specifies the search field.
When this field matches a search constant, the input field will be
changed in the output to a replacement value.
o The length of the output replacement field. Permissible length is 1 to
64 bytes.
srch The search constant to which the search field is compared. Permissi-
ble formats are character string (C'x...x'), hexadecimal string

Syncsort MFX Programmer’s Guide 2–183

OUTREC

(X'x...x'), or a binary byte (B'bbbbbbbb'). When the search constant

matches the search field, the input field will be changed to an output
replacement value.
If one of the search constants is binary in a set of search-replacement
pairs on a CHANGE subparameter, then all the search constants on
that CHANGE subparameter must be binary. (For additional infor-
mation on using binary fields in INCLUDE/OMIT processing, see
“INCLUDE/OMIT Control Statement” on page 2-29.)
If the search constant is longer than the length of the search field,
the constant will be truncated to the length of the search field. If the
search constant is shorter, the constant will be padded on the right.
Character strings are padded with blanks (X'40'). Hexadecimal
strings are padded with zeros (X'00'). Binary strings are neither
truncated nor padded since only one-byte strings are permissible.
repl The replacement value to which the input field is changed in the
reformatted output record when the search field matches a search
constant. The replacement value can either be a constant, a field
from the input record, or a fixed-length parsed field.
The term repl represents the following syntax:

 mrepl 
 j,k 
 %pp 

Figure 89. Syntax of repl

mrepl A replacement constant to which the input field is

changed. The replacement formats that are permissible
for constants are character string (C'x…x') and hexadeci-
mal string (X'x…x').
If the replacement constant is longer than the length o of
the output field, the constant will be truncated to length o.
If the replacement constant is shorter than o, the constant
will be padded on the right to length o. Character strings
are padded with blanks (X'40'). Hexadecimal strings are
padded with zeros (X'00').

j,k The position j and the length k of an input field that will
be inserted in the output record. k must be at least 1 and
cannot be greater than the length o specified for the out-
put replacement field. If k is less than o, the field j,k will
be padded on the right with blanks (X'40') to the length o.

2–184 Syncsort MFX Programmer’s Guide

 OUTREC

%pp Identifies a fixed-length parsed field that will be inserted

in the output record. The length of %pp specified by FIX-
LEN cannot be greater than the length o specified for the
output replacement field. If FIXLEN is less than o, the
%pp field will be padded on the right with blanks (X'40') to
the length o.
NOMATCH Indicates how MFX should respond if the input field does not match
a search constant. If NOMATCH is not specified and no search con-
stant matches the input field, sort processing will terminate with an
error message.
nmrepl A replacement constant to which the input field is changed in the
reformatted output record when the search field p,l fails to match a
search constant. For details, see the description of the repl variable
above.
r,n The position r and length n of an input field that will be inserted in
the output record when the CHANGE search field fails to match a
search constant.
n must be at least 1. If n is greater than the length o specified for the
output replacement field, the output field r,n will be truncated on the
right to length o. If n is less than o, the field r,n will be padded on the
right with blanks (X'40') to the length o.
%pp Identifies a fixed-length parsed field that will be inserted in the out-
put record when the CHANGE search field fails to match a search
constant. The length of %pp specified by FIXLEN cannot be greater
than the length o specified for the output replacement field. If FIX-
LEN is less than o, the %pp field will be padded on the right with
blanks (X'40') to the length o.
The following example illustrates the use of the CHANGE subparameter:

OUTREC FIELDS=(16,2,
CHANGE=(13,C'NJ',C'NEW JERSEY',
C'NY',C'NEW YORK',
C'PA',C'PENNSYLVANIA',
C'XX',50,13),
NOMATCH=(C'NOT SUPPORTED'),
8X,
24,1,
CHANGE=(10,B'1.......',C'EAST COAST',
B'0.......',C'WEST COAST'))

Figure 90. Sample OUTREC Parameter with CHANGE Subparameter

In the above example, the FIELDS parameter contains two CHANGE

subparameters. The first CHANGE subparameter changes the input field in
columns 1 through 13 to a state name in the reformatted output record when the

Syncsort MFX Programmer’s Guide 2–185

OUTREC

search field in column 16 matches a state code. If the state code is XX, positions 50
through 62 of the input record will be placed in the output record. If no matches are
found, the output field will be 'NOT SUPPORTED.' The second change
subparameter changes the one-byte input field in column 22 to 'EAST COAST' or
'WEST COAST' in the reformatted output record, depending on the binary contents
of the search field in column 24.
The following example illustrates a situation that can arise when using binary
search constants. In such cases, more than one search constant may match a search
field:

OUTREC FIELDS=(24,1,
CHANGE=(6,B'.....11.',C'SHARE',
B'.....01.',C'UNIQUE'))

Figure 91. CHANGE Subparameter with Binary Search Constants

Note that in the above example, the search field X'06' would match both binary
search constants. In such cases, the first search constant is used, thus the output
would be the character string 'SHARE'. If the search field were X'02', the output
would be the character string 'UNIQUE'.

JFY Subparameter
The JFY subparameter specifies that an input field be processed for left-
justification or right-justification for the output record. The JFY subparameter
specifies the following basic operations:
• If left-justification is specified, leading blank characters are eliminated; all
remaining characters are shifted left; if necessary, blank characters are
introduced to the right to create a fixed-length field.
• If right-justification is specified, trailing blank characters are eliminated;
all remaining characters are shifted right; if necessary, blank characters are
introduced to the left to create a fixed-length field.
• If a variable-length field is requested, all characters are shifted left and
trailing blanks are eliminated.
The JFY subparameter also can specify the following options:
• Introduce new leading and trailing nonblank characters.
• Eliminate previous leading and trailing nonblank characters.
• Change the length of the field in the output record.
The format of the JFY subparameter is shown below:

2–186 Syncsort MFX Programmer’s Guide

 OUTREC

 
 p,l,   SHIFT=LEFT 
[c:]   JFY=(  SHIFT=RIGHT  [ ,LENGTH=n ] [ ,PREBLANK=list ]
 %pp,   VL 
 

[ ,LEAD=string ] [ ,TRAIL=string ] )

Figure 92. JFY Subparameter

The following describes the elements of the JFY subparameter:
p,l Specifies the beginning byte position p and byte length l of the
input record’s relevant field.
%pp Specifies a fixed-length parsed field. See “PARSE Parameter
(Optional)” on page 2-209 for further description.
SHIFT=LEFT Specifies left-justification of the input field. Leading blank
characters are eliminated; all remaining characters are
shifted left; any necessary blanks are introduced on the right
side to compensate for the length of the output field. (The
default output field length is equal to the input field length.)
SHIFT=RIGHT Specifies right-justification of the input field. Trailing blank
characters are eliminated; all remaining characters are
shifted right; any necessary blanks are introduced on the left
side to compensate for the length of the output field. (The
default output field length is equal to the input field length.)
VL Specifies that a variable-length field should be produced. The
field will be left-justified and trailing blanks will be deleted.
The output record must be a variable-length record.
The maximum length of the field is determined from the
input field length and the lengths of any LEAD and TRAIL
strings.
Any INREC/OUTREC fields following a VL field cannot spec-
ify a starting column number or any of the alignment options
(H, F or D for halfword, fullword or doubleword alignment).
Also, VL cannot be used with the OVERLAY parameter or the
LENGTH subparamter.
VL is permitted with fixed-length data and OUTFIL FTOV. It
is also permitted with OUTFIL IFTHEN and FTOV, but not
with a WHEN=INIT parameter or with a HIT=NEXT param-
eter when FTOV is used.

Syncsort MFX Programmer’s Guide 2–187

OUTREC

LENGTH=n Optionally alters the length of the output field to accommo-

date a change in the total number of characters from the
input field. The default output field length is equal to the
input field length l. Use LENGTH=n to either extend a field
that needs to be larger due to the addition of a leading or
trailing string, or to shorten a field from the default length to
reduce the number of padding blank characters. LENGTH
cannot be specified with the VL subparameter.
PREBLANK=list Optionally specifies nonblank leading or trailing characters to
be replaced with blank characters before justification. The
characters to be replaced are specified together in a character
string constant (C'string') or hexadecimal string constant
(X'hh...hh') from 1 to 10 bytes. PREBLANK searches from left
to right for leading characters, and from right to left for trail-
ing characters. A blank character is substituted for each indi-
vidual leading or trailing character that matches any
individual PREBLANK string character. The search termi-
nates for each leading and trailing side when the first non-
blank character not on the list is encountered. For example,
PREBLANK=C'<>' replaces each occurrence of '<' and '>' with
a blank character in the leading and trailing characters
before justification.
LEAD=string Optionally specifies a character constant (C'string') or hexa-
decimal constant (X'hh...hh') from 1 to 50 bytes that is placed
in the output field immediately left of the first nonblank char-
acter in the field. Note that LENGTH=n also may need to be
specified to accommodate the additional leading characters.
For example, LEAD=C'(' inserts '(' as a leading character in
the field.
TRAIL=string Optionally specifies a character constant (C'string') or hexa-
decimal constant (X'hh...hh') from 1 to 50 bytes that is placed
in the output field immediately right of the last nonblank
character in the field. Note that LENGTH=n also may need to
be specified to accommodate the additional trailing charac-
ters. For example, TRAIL=C')' inserts ')' as a trailing charac-
ter in the field.
The following example illustrates the use of the JFY subparameter:

11,18,JFY=(SHIFT=LEFT,LENGTH=15,PREBLANK=C'/[]*{}',
LEAD=C'<',TRAIL=C'>')

Figure 93. Sample JFY Subparameter

2–188 Syncsort MFX Programmer’s Guide

 OUTREC

In the example above, the field (11,18) is specified for left-justification for the
output record; LENGTH changes the output length to 15 bytes to reduce the output
field size; PREBLANK substitutes a blank character for each instance of /, [, ], *, {
and } in the leading and trailing characters before justification; LEAD introduces <
to the left of the first nonblank character in the justified field; and TRAIL
introduces > to the right of the last nonblank character in the justified field.
The following example illustrates the use of JFY with the VL subparameter:

OUTFIL OUTREC=(5,40,JFY=(VL),C’,’,45,60,JFY=(VL),C’,’,105,96,
JFY=(VL)),FTOV

Figure 94. Sample JFY with VL Subparameter

In the example above, a fixed-length record file is changed to a shorter variable-

length record file. The 40-byte field starting in position 5, the 60-byte field starting
in position 45 and the 96-byte field starting in position 105 are converted to
variable-length fields by removing leading and trailing blanks, and commas
separate the fields. Note that FTOV must be specified to create a variable-length
record output file.

SQZ Subparameter
The SQZ subparameter specifies that an input field be processed for “left-
squeezing” or “right-squeezing” for the output record. It includes the justification
functions of the JFY subparameter but adds elimination of all blank characters in
the input field and additional options for selecting and replacing blank and
nonblank characters.
The SQZ subparameter specifies the following basic operations:
• All blank characters in the input field are eliminated.
• If left-shifting is specified, the remaining characters are shifted left; any
necessary blanks are introduced on the right to create a fixed-length field.
• If right-shifting is specified, the remaining characters are shifted right; any
necessary blanks are introduced on the left to create a fixed-length field.
• If a variable-length field is requested, all characters are shifted left and
trailing blanks are eliminated.
The SQZ subparameter also can specify the following options:
• Introduce leading and trailing nonblank characters.
• Replace user-specified nonblank characters with blank characters prior to
squeeze operation.
• Replace blank characters with user-specified nonblank characters.
• Retain blank characters between paired apostrophes.
• Retain blank characters between paired quotes.
• Change the length of the field in the output record.

Syncsort MFX Programmer’s Guide 2–189

OUTREC

The format of the SQZ subparameter is shown below:

 
 p,l,   SHIFT=LEFT 
[c:]   SQZ=(  SHIFT=RIGHT  [ ,LENGTH=n ] [ ,PREBLANK=list ]
 %pp,   VL 
 

[ ,LEAD=string ] [ ,MID=string ] [ ,TRAIL=string ]  ,PAIR=APOST  )

 ,PAIR=QUOTE 

Figure 95. SQZ Subparameter

The following describes the elements of the SQZ subparameter:
p,l Specifies the beginning byte position p and byte length l of the
input record’s relevant field.
%pp Specifies a fixed-length parsed field. See “PARSE Parameter
(Optional)” on page 2-209 for further description.
SHIFT=LEFT Specifies left-squeezing of the input field. By default, all
blank characters are eliminated and all nonblank characters
are shifted left; if necessary, blank characters are introduced
on the right side to compensate for the length of the output
field. (The default output field length is equal to the input
field length.)
SHIFT=RIGHT Specifies right-squeezing of the input field. By default, all
blank characters are eliminated and all nonblank characters
are shifted right; if necessary, blank characters are intro-
duced on the left side to compensate for the length of the out-
put field. (The default output field length is equal to the input
field length.)
VL Specifies that a variable-length field should be produced. The
field will be left-justified and all blank characters will be
deleted. The output record must be a variable-length record.
The maximum length of the field is determined from the
input field length and the lengths of any LEAD, MID and
TRAIL strings.
Any INREC/OUTREC fields following a VL field cannot spec-
ify a starting column number or any of the alignment options
(H, F or D for halfword, fullword or doubleword alignment).
Also, VL cannot be used with the OVERLAY parameter or the
LENGTH subparamter.
VL is permitted with fixed-length data and OUTFIL FTOV. It
is also permitted with OUTFIL IFTHEN and FTOV, but not

2–190 Syncsort MFX Programmer’s Guide

 OUTREC

with a WHEN=INIT parameter or with a HIT=NEXT param-

eter when FTOV is used.
LENGTH=n Optionally alters the length of the output field to accommo-
date a change in the total number of characters from the
input field. The default output field length is equal to the
input field length l. Use LENGTH=n to either extend a field
that needs to be larger due to the addition of a leading or
trailing string, or to shorten a field from the default length to
reduce the number of padding blank characters. LENGTH
cannot be specified with the VL subparameter.
PREBLANK=list Optionally specifies user-defined nonblank characters to be
replaced with blank characters before squeezing. The charac-
ters to be replaced are specified together in a character string
constant (C'string') or hexadecimal string constant
(X'hh...hh') from 1 to 10 bytes. PREBLANK searches through-
out the entire input field and substitutes a blank character for
each individual nonblank character that matches any indi-
vidual PREBLANK string character. For example, PREB-
LANK=C'<>' replaces each occurrence of '<' and '>' with a
blank character throughout the entire input field before
squeeze operation.
LEAD=string Optionally specifies a character constant (C'string') or hexa-
decimal constant (X'hh...hh') from 1 to 50 bytes that is placed
in the output field immediately left of the first nonblank char-
acter in the field. Note that LENGTH=n also may need to be
specified to accommodate the additional leading characters.
For example, LEAD=C'(' inserts '(' as a leading character in
the field.
MID=string Optionally specifies insertion of a string per one or group of
adjoining blank characters that is eliminated between the
first and last nonblank characters in the field. The string can
be a character constant (C'string') or hexadecimal constant
(X'hh...hh') from 1 to 10 bytes. Note that LENGTH=n also
may need to be specified to accommodate the additional char-
acters. For example, MID=C'*' substitutes a '*' for every group
of one or more adjoining blank characters that have been
eliminated between the first and last nonblank characters in
the field.
TRAIL=string Optionally specifies a character constant (C'string') or hexa-
decimal constant (X'hh...hh') from 1 to 50 bytes that is placed
in the output field immediately right of the last nonblank
character in the field. Note that LENGTH=n also may need to
be specified to accommodate the additional trailing charac-

Syncsort MFX Programmer’s Guide 2–191

OUTREC

ters. For example, TRAIL=C')' inserts ')' as a trailing charac-

ter in the field.
PAIR=APOST Optionally specifies that all blank and PREBLANK charac-
ters between pairs of apostrophes remain unchanged. Any
apostrophe throughout the field also remains unchanged. For
an unpaired apostrophe, if SHIFT=LEFT or VL is specified,
the characters are unchanged from the apostrophe rightward
to the end of the field; if SHIFT=RIGHT is specified, the char-
acters are unchanged from the apostrophe leftward to the
beginning of the field.
PAIR=QUOTE Optionally specifies that all blank and PREBLANK charac-
ters between pairs of quotes remain unchanged. Any quote
throughout the field also remains unchanged. For an
unpaired quote, if SHIFT=LEFT or VL is specified, the char-
acters are unchanged from the quote rightward to the end of
the field; if SHIFT=RIGHT is specified, the characters are
unchanged from the quote leftward to the beginning of the
field.
The following example illustrates the use of the SQZ subparameter:

11,18,SQZ=(SHIFT=LEFT,LENGTH=30,PREBLANK=C'/[]*{}',
LEAD=C'<',MID=C',',TRAIL=C'>',PAIR=QUOTE)

Figure 96. Sample SQZ subparameter

In the example above, the field (11,18) is specified for left-squeezing for the output
record; LENGTH changes the output length to 30 bytes to accommodate added
characters; PREBLANK substitutes a blank character for each instance of /, [, ], *, {
and } throughout the entire input field before squeezing; LEAD introduces < to the
left of the first nonblank character in the squeezed field; MID substitutes a comma
for each group of blank characters between the first and last nonblank characters;
TRAIL introduces > to the right of the last nonblank character in the squeezed
field; and PAIR specifies that all blank and PREBLANK characters between pairs
of quotes remain unchanged.

CONVERT Parameter (Optional)

The CONVERT parameter enables you to convert variable-length records into
fixed-length records.
These records do not require an RDW and will be written to any output file(s) with
a RECFM of F or FB. When using CONVERT, you no longer need to apply the rules
for “Specifying the FIELDS parameter for Variable-Length Records.”
You cannot specify the variable portion of the input records (position without
length) when using CONVERT. However, all data fields need not be present in each

2–192 Syncsort MFX Programmer’s Guide

 OUTREC

record being CONVERTed, unless a numeric or year data field is specified. That is,
blanks will be used as a default for any missing p,l field bytes, while all p,l,f fields
must be present.
When using CONVERT in conjunction with the OUTREC parameter on the
OUTFIL control statement, data fields of any type need not be present, and you
may change the default padding character with the VLFILL parameter. (See the
explanations of CONVERT and VLFILL in the OUTFIL control statement section.)
You may also create multiple output files with different record formats when
specifying CONVERT on the OUTFIL control statement.

VTOF Parameter (Optional)

VTOF is equivalent to CONVERT. See “CONVERT Parameter (Optional)” on
page 2-192.

FINDREP Parameter (Optional)

The FINDREP parameter allows you to find one or more constants in a record and
replace them with a provided constant.
FINDREP compares the current position in an input record to an input constant,
seeking a match. By default it starts with position 1 for fixed-length records and
position 5 for variable-length records. The current position will increase by 1 until
a match is found. Once FINDREP discovers a match at the current position, the
output constant will supplant the input constant, the current position will advance
beyond the location of the replaced input constant, and the process will continue.
Bytes appearing after the replaced constants will be moved either left or right as
necessary until the current position reaches the record's final position and
processing ceases.
For fixed-length records if a record needs to be shortened due to a shorter replace
constant, it will be padded with trailing blanks as needed. If a fixed-length record
is lengthened, trailing blank characters will be removed. Variable-length records
will have their length adjusted as appropriate. If a variable-length record exceeds
its maximum record length, trailing blanks will be deleted.
FINDREP processing requires input and output constants to be specified. An input
constant can be any of the following: a single hexadecimal string, a repeated
hexadecimal string, a single character string, or a repeated character string. An
output constant can be any of the following: a single hexadecimal string, a repeated
hexadecimal string, a single character string, a repeated character string, or a null
string. Permissible syntax expressions for input and output constants are listed
below.
• C’string’
• nC’string’
• X’string’
• nX’string’

Syncsort MFX Programmer’s Guide 2–193

OUTREC

• C” This expression can only be used to define a null output constant.

Note the following considerations for defining input and output constants:
• The maximum length of either an input or an output constant is 256 bytes.
• Two apostrophes must be used to specify a single apostrophe.
• Input constants can be removed through use of a null output constant.
The format of the FINDREP parameter is shown below:

)
  IN=(ic 1 [ ,ic 2 ]… [ ,ic n ] ) ,OUT=oc  
 [ ,STARTPOS=p ] [ ,ENDPOS=q ]
  INOUT=(ic 1 ,oc 1 [ ,ic 2 ,oc 2 ]… [ ,ic n ,oc n ] )  
 
FINDREP=(  
  ERROR   YES  
 [ , DO=n ] [ ,MAXLEN=m ] , OVERRUN = -------------------
 TRUNC  - , SHIFT = ----------
 NO   -
     

Figure 97. FINDREP Parameter Format

IN=(ic1[,ic2]…[,icn] ) Specifies one or more input constants that will be searched

for during the FINDREP operation. Each ic specifies an input
constant to search for. See description above on how to specify
input constants.
OUT=oc Specifies the output constant that will be used to replace any
of the input constants that are found. oc represents the out-
put constant used in the replace operation. See description
above on how to specify output constants.
INOUT=(ic1,oc1[,ic2,oc2]…[,icn,ocn] Specifies pairs of input constants and out-
put constants that will be used in the FINDREP operation.
Each ic specifies an input constant to search for. Each oc rep-
resents the output constant for the replace operation. See
description above on how to specify input and output con-
stants.
By default, the FINDREP function starts at position 1 for fixed-length records or
position 5 for variable-length records and ends processing at the end of the record.
The following options will alter the default FINDREP behavior:
STARTPOS=p Use this option to change the starting position of a fixed-
length record’s default position of 1 or a variable-length
record’s default position of 5. The STARTPOS=p option uses
the variable p to denote starting position. For variable-length
records, p will be reset to 5 if a value less than 5 is specified. If
p is greater than the length of the input record, FINDREP
will perform no action on the record.

2–194 Syncsort MFX Programmer’s Guide

 OUTREC

ENDPOS=q Use this option to change the ending position of the FIND-
REP operation. The ENDPOS=q option uses the variable q to
denote the last position to scan in the input record. For vari-
able-length records, q will be reset to 5 if a value less than 5 is
specified. When both ENDPOS=q and STARTPOS=p are
defined, if q is less than p, FINDREP will perform no action
for the record. ENDPOS does not affect the shifting of bytes
during the FINDREP operation.
DO=n Use this option to limit the maximum number of times FIND-
REP will be performed for a record. The DO=n option uses the
n variable to denote the number of times an input constant is
found and replaced, in which n can be between 1 and 1000.
FINDREP will stop scanning for input constants when n
input constants have been found and replaced.
MAXLEN=m Use this option to change the maximum length of the output
record, which has a default of the maximum record length
input to FINDREP. The MAXLEN=m option uses the m vari-
able to denote the maximum length of the record. MAXLEN
can increase or decrease the record length, except for an IFT-
HEN FINDREP where it can only increase the record length.
MAXLEN can be used to increase the record length when the
replace constants are longer than the find constants. For fur-
ther details, see the "Considerations" section.
OVERRUN=ERROR Use the default OVERRUN=ERROR option to specify how
overruns are handled by MFX. Overruns will occur when non-
blank bytes need to be shifted beyond the maximum record
length or when MAXLEN=n is used to shorten an output
record’s length to be fewer than the total number of trailing
non-blank bytes. OVERRUN=ERROR will issue the
WER439A error message and terminate MFX when overruns
occur.
OVERRUN=TRUNC Use the OVERRUN=TRUNC option to truncate the output
record and prevent an error message from appearing if an
overrun happens. If you choose to employ the OVER-
RUN=TRUNC option, MFX will eliminate all bytes beyond
the end of the output record length.
 YES 
SHIFT =  -----------  Use this option to change how an output constant will replace
 NO 
an input constant of a different length. The default
SHIFT=YES option will instruct MFX to accommodate longer
output constants by shifting bytes to the right and shorter
output constants by shifting bytes to the left. If you select the
SHIFT=NO option, MFX will overlay an input constant with
its corresponding output constant without moving bytes left
or right. When SHIFT=NO is specified, the current position

Syncsort MFX Programmer’s Guide 2–195

OUTREC

for FINDREP will be advanced by the shorter of the input and

output constants when a match is found.
Considerations:
• In an IFTHEN clause, you cannot use FINDREP with BUILD or OVERLAY.
• In an INREC or OUTREC statement, you cannot use FINDREP with BUILD,
OVERLAY, IFTHEN or IFOUTLEN, although FINDREP within an IFTHEN
clause is permitted.
• In an OUTFIL statement, you cannot use FINDREP with BUILD, OVERLAY,
IFTHEN, IFOUTLEN, VTOF, CONVERT or VLFILL, although FINDREP
within an IFTHEN clause is permitted.
• After a constant has been replaced at the current position in a single FINDREP
option, no further checks are performed at that position. One FINDREP
statement cannot be used to replace a constant and then replace the original
constant’s replacement.
• In an IFTHEN clause employing FINDREP on a fixed-length record, the
FINDREP uses the input record length.
The following examples outline the proper uses of the FINDREP parameter and its
subparameters.

Example 1
The following is an example of the FINDREP parameter on an OUTREC statement
which is used to replace state abbreviations in a record with their full names. The
FINDREP operation is restricted to columns 60 to 75. The input consists of 80-byte
fixed-length records. Trailing blanks will be truncated from the records when the
replace constant is substituted.

OUTREC FINDREP=(INOUT=(C’, NJ’,C’, NEW JERSEY’,C’, NY’,

C’, NEW YORK’),STARTPOS=60,ENDPOS=75)

Figure 98. Sample FINDREP Parameter

If the original records contained:

Col 1..........................col 60 80
.. WOODCLIFF LAKE, NJ 07677
...... ..NEW ROCHELLE, NY 10801

Figure 99. Sample of Original Records

2–196 Syncsort MFX Programmer’s Guide

 OUTREC

The modified records would contain:

Col 1..........................col 60 80
.. WOODCLIFF LAKE, NEW JERSEY 07677
.......NEW ROCHELLE, NEW YORK 10801

Figure 100. Sample of Modified Records

In the above example, if the values to be replaced were positioned closer to column
80, it would be possible that the trailing zip code would be pushed past column 80.
By default this would generate an error since only blanks can be truncated from a
record. This error could be avoided by specifying the MAXLEN parameter to extend
the record length to accommodate the new longer replacement literals. The
following OUTREC control statement could address this problem:

OUTREC FINDREP=(INOUT=(C’, NJ’,C’, NEW JERSEY’,C’, NY’,

C’, NEW YORK’),STARTPOS=60,ENDPOS=80,MAXLEN=90)

Figure 101. Sample FINDREP Parameter with MAXLEN

If non-blank characters after the replacement string are not needed, the
OVERRUN=TRUNC option can be specified to remove the trailing characters from
the record. For example, the following can be used:

OUTREC FINDREP=(INOUT=(C’, NJ’,C’, NEW JERSEY’,C’, NY’,

C’, NEW YORK’),STARTPOS=60,ENDPOS=80,OVERRUN=TRUNC)

Figure 102. Sample FINDREP Parameter with OVERRUN=TRUNC

Example 2
In the following example an INREC statement will be used to abbreviate each
instance of ‘NEW JERSEY’ and ‘NEW YORK’ in a record when position 24 of the
record contains a X’01’.

INREC IFTHEN=(WHEN=(24,1,BI,EQ,X’01’),
FINDREP=(INOUT=(C’NEW JERSEY’,C’NJ’,C’NEW YORK’,
C’NY’)))

Figure 103. Sample FINDREP Parameter

For the input record:

NEW YORK,ABC NEW JERSEY,XYZ,NEW YORK

Figure 104. Sample Input Record

Syncsort MFX Programmer’s Guide 2–197

OUTREC

The output record would contain:

NY,ABC NJ,XYZ,NY

Figure 105. Sample Output Record

If the records are variable-length, the RDW of the record would be reduced to
indicate the new length after the shorter literals are substituted. If the records are
fixed-length, spaces would be appended to the end of the record to replace the
deleted characters.

Example 3:
In the following example, the input constant will be replaced by the output
constant without shifting bytes to accommodate the replaced constant. This can be
used to replace only a portion of the input search constant with a new constant. In
this example, a portion of the input constant will be replaced.

OUTREC FINDREP=(IN=(C’CODE=VALID’,C’CODE=INVALID’),
OUT=C’FLAG’,SHIFT=NO)

Figure 106. Sample FINDREP Parameter

For the input record:

CODE=VALID ABC 123 CODE=INVALID CODE=UNKNOWN

Figure 107. Sample Input Record

The output record would be

FLAG=VALID ABC 123 FLAG=INVALID CODE=UNKNOWN

Figure 108. Sample Output Record

Note that in this instance where the output constant is shorter than the input
constant, the FINDREP operation will resume at the next character after the
output constant rather than the input constant as would normally be the case.

IFTHEN Parameter (Optional)

The IFTHEN parameter employs conditional logic, which enables you to reformat
your records based on specified criteria. Multiple IFTHEN parameters may be
specified within the same control statement and are processed sequentially.
The IFTHEN parameter may be used within the INREC, OUTREC, and OUTFIL
control statements.

2–198 Syncsort MFX Programmer’s Guide

 OUTREC

The format of the IFTHEN parameter is illustrated below.

  
   ,BUILD= ( fields )   
 
 [ ,PARSE=(subparm) ]  ,OVERLAY= ( fields )    
WHEN=INIT   
   ,FINDREP=(subparm)   
    
  ,PARSE=(subparm)  
 
WHEN=GROUP,selectoption[,selectoption][,selectoption], PUSH=(pushsubparm) 
where selectoption is: 
 BEGIN=(conditions) 
 
 END=(conditions) 
 KEYBEGIN=(p,l) 
 RECORDS=n 
 
   
IFTHEN= (     )
 [,PARSE=(subparm)]  ,BUILD  =(fields) 
WHEN=(conditions)  ,OVERLAY  [,HIT=NEXT] 
  
 
 ,FINDREP=(subparm)  
 
 
 
  ,BUILD  
WHEN=ANY [,PARSE=(subparm)]   =(fields) 
 ,OVERLAY  [,HIT=NEXT]
 
 ,FINDREP=(subparm) 
 
 
   
 ,BUILD 
WHEN=NONE [ ,PARSE=(subparm) ]   =(fields) 
 ,OVERLAY  [,HIT=NEXT]
 
 ,FINDREP=(subparm) 

Figure 109. IFTHEN Parameter Format

At the beginning of IFTHEN processing, a temporary record is created from each of

your input records.
The IFTHEN parameter automatically makes the following changes to the
temporary record to accommodate any adjustments in length. In a variable-length
record, the RDW length is adjusted accordingly. In a fixed-length record, the record
is padded with blanks when necessary. Blanks also replace missing bytes in input
fields.
The IFTHEN parameter has two main parts: the WHEN subparameter and a
second subparameter. As shown in Figure 109 on page 2-199, the WHEN
subparameter may be WHEN=INIT, WHEN=GROUP, WHEN=(conditions),
WHEN=ANY, or WHEN=NONE. Except for WHEN=GROUP, the second
subparameter may be FINDREP, PARSE and BUILD or OVERLAY. The WHEN
subparameter defines a condition that must be satisfied before the second
subparameter is applied to the temporary records. If the WHEN subparameter
condition is not satisfied, then the second subparameter is not applied to the
temporary records.
Since IFTHEN parameters refer to the temporary records instead of the input
records, all subsequent IFTHEN parameters within the same control statement
will take previous FINDREP, BUILD or OVERLAY changes into account. Once

Syncsort MFX Programmer’s Guide 2–199

OUTREC

IFTHEN processing for each record stops, the temporary record becomes the output
record.
When SEQNUM is used within a BUILD or OVERLAY parameter in an IFTHEN
clause, the sequence number will be incremented each time that the BUILD or
OVERLAY for that clause is performed. This may lead to different sequence
numbers being generated for the same input record when SEQNUM is used in
different IFTHEN clauses.
You can use %pp parsed fields in IFTHEN expressions. If the %pp field is defined in
a WHEN=INIT, WHEN=(conditions), WHEN=ANY, or WHEN=NONE expression,
it can be used in the IFTHEN BUILD or IFTHEN OVERLAY of that expression.
Additionally, for WHEN=INIT, the %pp fields can be used in any subsequent
IFTHEN BUILD or OVERLAY expression. See “PARSE Parameter (Optional)” on
page 2-209 for further description on PARSE.
The following describes the IFTHEN subparameters:
WHEN=INIT The WHEN=INIT subparameter condition is automatically
satisfied. It applies the remaining IFTHEN subparameters to
each temporary record.
A second subparameter is required. PARSE is optional if
BUILD or OVERLAY is specified.
WHEN=GROUP The WHEN=GROUP subparameter is satisfied if a record
meets the specified grouping options. A WHEN=GROUP
clause can be combined with WHEN=INIT clauses, but must
be defined before using the WHEN=(conditions),
WHEN=NONE, or WHEN=ANY subparameters.
WHEN=GROUP groups records and propagates fields, identi-
fiers, and sequence numbers based on the criteria specified in
the BEGIN=(conditions), END=(conditions), KEYBE-
GIN=(p,l), RECORDS=n, and PUSH=(c:item,...) options. At
least one BEGIN=(conditions), END=(conditions), KEYBE-
GIN=(p,l) or RECORDS=n option must be specified.
BEGIN=(conditions) Determines the logical test used to
specify that a record starts a group. Each
record meeting the criteria set by the logical
test will begin a new group.
Under the INCLUDE/OMIT control state-
ment, see “COND Parameter (Required)” on
page 2-31 for a complete description of com-
parisons and logical expressions. However,
the following cannot be used in
WHEN=GROUP:
•D2 format
•FORMAT=f

2–200 Syncsort MFX Programmer’s Guide

 OUTREC

•Locale processing
•VLTESTI (during IFTHEN processing,
blanks replace missing bytes in input fields)
END=(conditions) Determines the logical test used to spec-
ify that a record ends a group. Each record
meeting the criteria set by the logical test
will end a group. All logical expressions
used for the BEGIN option, discussed
above, can be used for the END option.
KEYBEGIN=(p,l) Establishes the start of a new group for a
record when the field in the record begin-
ning in column p for length l changes. The
first input record will start a group. The
maximum column p is 32752 and the maxi-
mum length l is 256. If KEYBEGIN and
BEGIN are both used in a WHEN=GROUP
clause, a new group will begin when either
parameter dictates one. A dictionary_name
may be used for p,l.
RECORDS=n Determines the maximum number of
records, defined by n, that can be contained
in a group. This number can be defined
from 1 to 2000000000. If none of the KEY-
BEGIN, BEGIN, or END options is specified
when a RECORDS option is defined, every
n records will be grouped together.
PUSH=([c:]psh1[,psh2]... [,pshn]) Defines the input field,
sequence number, or identifier that will be
overlaid for each group’s records. The fol-
lowing options can be used to define PUSH:
c: Specifies a record’s output column that
will be overlaid. If c: is not defined for the
first item, 1 will be used as a default. For
variable-length records, column 5 or higher
should be specified to avoid overlaying the
RDW. If c: is not specified for any item, the
next item starts immediately after the pre-
vious item.
If the value used for c extends the output
record beyond the input record, blank bytes
will be added to the left, increasing the
record’s length. Should the c value extend
the length of a variable-length record, the

Syncsort MFX Programmer’s Guide 2–201

OUTREC

RDW length will be adjusted after all of the

items are processed.
The following describes the psh elements
that can be placed in a record:
p,l Specifies the position and length of a
field to propagate from each group’s first
input record to every record in the group.
Blanks will replace missing bytes within
specified input fields, allowing the short or
missing fields to be processed.
ID=n Specifies a printable Zoned Decimal
(ZD) identifier n bytes long, which will be
added to every record of each group. For the
first group, the identifier will start at 1 and
for each subsequent group it will be
increased by 1. The number n can be from 1
to 15.
SEQ=n Specifies a printable ZD sequence
number n bytes long, which will be added to
every record of each group. For the first
record of each group, the identifier will start
at 1 and for each subsequent record it will
be increased by 1. The number n can be
from 1 to 15.
WHEN=(conditions) The WHEN=(conditions) subparameter condition is satis-
fied if a temporary record meets the specified conditions. It
applies the specified second subparameter to each temporary
record that meets the specified conditions.
conditions The conditions must be formulated into a com-
parison or logical expression that can be evalu-
ated to true or false.
Under the INCLUDE/OMIT control statement,
see “COND Parameter (Required)” on page 2-31
for a complete description of comparisons and
logical expressions. However, the following can-
not be used in WHEN=(conditions):
• D2 format
• FORMAT=f
• Locale processing

2–202 Syncsort MFX Programmer’s Guide

 OUTREC

• VLTESTI (during IFTHEN processing,

blanks replace missing bytes in input
fields)
The second subparameter is required. The PARSE and
HIT=NEXT subparameters are optional.
WHEN=ANY The WHEN=ANY subparameter condition is satisfied if one
or more of its associated WHEN=(conditions) subparameter
conditions have been satisfied. Its associated WHEN=(condi-
tions) subparameters are those that precede it but no other
WHEN=ANY subparameter. If the WHEN=ANY subparame-
ter condition is satisfied, it applies the specified second sub-
parameter to the temporary record.
The second subparameter is optional. If it is not used, IFT-
HEN processing simply stops if the WHEN=ANY subparame-
ter condition is satisfied unless the optional HIT=NEXT
subparameter has been specified.
WHEN=NONE The WHEN=NONE subparameter condition is satisfied if
none of the preceding WHEN=(conditions) subparameter con-
ditions is satisfied or if there are no WHEN=(conditions) sub-
parameters. If the WHEN=NONE subparameter condition is
satisfied, it applies the specified second subparameter to the
temporary record.
The second subparameter is optional. If it is not used, IFT-
HEN processing simply stops if the WHEN=NONE subpa-
rameter condition is satisfied unless the optional HIT=NEXT
subparameter has been specified.
The IFTHEN parameters must be specified such that the WHEN subparameters
are in the following order:
• WHEN=INIT and/or WHEN=GROUP
• WHEN=(conditions) and WHEN=ANY
• WHEN=NONE
BUILD Except for the WHEN=GROUP subparameter, the IFTHEN
parameter will accept BUILD as a second subparameter. See
“FIELDS/BUILD Parameter” on page 2-137 for a complete
description of the BUILD subparameter.
OVERLAY Except for the WHEN=GROUP subparameter, the IFTHEN
parameter will accept OVERLAY as a second subparameter.
See “OVERLAY Parameter (Optional)” on page 2-208 for a
complete description of the OVERLAY subparameter.
HIT=NEXT The HIT=NEXT subparameter is optional, but can only be
used in conjunction with the WHEN=(conditions) or

Syncsort MFX Programmer’s Guide 2–203

OUTREC

WHEN=ANY subparameter. IFTHEN processing stops by

default once a WHEN=(conditions) subparameter condition or
WHEN=ANY subparameter condition is satisfied. Including
the HIT=NEXT subparameter will continue IFTHEN process-
ing regardless of whether or not the WHEN subparameter
condition is satisfied.
PARSE The PARSE parameter is optional except in WHEN=INIT if
BUILD or OVERLAY is not specified. PARSE is used to
extract variable-position and variable-length fields from
records and place the resultant data into fixed-length parsed
fields. PARSE cannot be used in a WHEN=GROUP clause.
See “PARSE Parameter (Optional)” on page 2-209 for further
description.
FINDREP The FINDREP parameter provides the ability to find and
replace one or more constants in a record. A constant to be
searched for can be specified as a character or hexadecimal
string and its replacement constant can be either a character,
hexadecimal or null string. Depending on the length of the
replacement constant, subsequent characters will be shifted
left or right. For fixed-length records, if data is shifted left,
the record will be padded with blanks as needed. If data is
shifted right, any trailing blank characters will be removed.
Variable-length records will have their length adjusted as
appropriate. If a variable-length record exceeds its maximum
record length, trailing blanks will be deleted.
Optionally, controls are provided to specify the positions to be
scanned, the number of times a find/replace operation can
occur, actions to be taken if a non-blank character needs to be
shifted past the record length, a new record length or whether
a replace or overlay of the find constant is to be performed.
See “FINDREP Parameter (Optional)” on page 2-193 for
details on its use.
The following example outlines the use of the WHEN=GROUP parameter.
WHEN=GROUP can be useful for keeping together groups of unlike input records,
enabling them to be correctly sorted.
For instance, if each transaction at a store generates a series of unlike records, and
all the records are collected in a file that needs to be sorted by date and register
number, WHEN=GROUP can generate appropriate sort keys for each record in the
group.
A header record (code ‘H’) with a register number and date, detail SKU records
(code ’S’) with an SKU number, unit price and quantity, and a trailer total record
(code ’T’) are generated for each transaction. A file with 20-byte fixed-length
records for three transactions might look like:

2–204 Syncsort MFX Programmer’s Guide

 OUTREC

H 0003 2008/08/17
S 872567 0010.22 001
S 510945 0001.99 003
S 734018 0003.98 002
T 0024.15
H 0005 2008/08/16
S 013298 0000.69 004
S 510945 0017.03 001
T 0019.79
H 0002 2008/08/17
S 212134 0003.49 003
T 0010.47
INREC WHEN=GROUP can be used with BEGIN to identify a header record
starting a group and END to identify a trailer record ending a group. PUSH
extends each record by placing the date and register number from the header
record at the end of each record in the group, followed by a 5-byte group number
and a 3-byte record sequence number. This enables all the records in a group to be
sorted together.

INREC IFTHEN=(WHEN=GROUP,BEGIN=(1,1,CH,EQ,C’H’),
END=(1,1,CH,EQ,C’T’),
PUSH=(21:8,10,31:3,4,35:ID=5,SEQ=3))

Figure 110. Sample WHEN=GROUP Parameter

The data will be transformed into:

H 0003 2008/08/17 2008/08/17000300001001

S 872567 0010.22 0012008/08/17000300001002
S 510945 0010.99 0032008/08/17000300001003
S 734018 0003.98 0022008/08/17000300001004
T 0024.15 2008/08/17000300001005
H 0005 2008/08/16 2008/08/16000500002001
S 013298 0000.69 0042008/08/16000500002002
S 510945 0017.03 0012008/08/16000500002003
T 0019.79 2008/08/16000500002004
H 0002 2008/08/17 2008/08/17000200003001
S 212134 0003.49 0032008/08/17000200003002
T 0010.47 2008/08/17000200003003

The records are then sorted using the new PUSH data.

SORT FIELDS=(21,10,CH,A,31,4,CH,A,35,8,CH,A)

Figure 111. Sample SORT Statement

The data added by PUSH can be eliminated from the output data with a simple
OUTREC statement for the original 20 bytes of the record, or by specifying

Syncsort MFX Programmer’s Guide 2–205

OUTREC

LRECL=20 on the SORTOUT DD statement, creating the following correctly sorted

output.

H 0005 2008/08/16
S 013298 0000.69 004
S 510945 0017.03 001
T 0019.79
H 0002 2008/08/17
S 212134 0003.49 003
T 0010.47
H 0003 2008/08/17
S 872567 0010.22 001
S 510945 0001.99 003
S 734018 0003.98 002
T 0024.15
If desired, a simple report can be created using OUTFIL IFTHEN to identify each
different record type, format it appropriately, and remove the data added by PUSH.
SECTIONS is used to generate a report header for each transaction.

OUTFIL SECTIONS=(35,5,HEADER3=(’ DATE REG# ID’),SKIP=L),

IFTHEN=(WHEN=(1,1,CH,EQ,C’H’), FOR HEADER RECORDS:

BUILD=(21,10,2X,3,4,2X,3,4,2X,35,5)), DATE, REG NUM, TRAN ID

IFTHEN=(WHEN=(1,1,CH,EQ,C’S’),

BUILD=(C’SKU#: ’,3,6,C’ PRICE: $’,10,7,

C’ QUANTITY: ’,18,3)),

IFTHEN=(WHEN=(1,1,CH,EQ,C’T’), FOR TRAILER RECORDS:

BUILD=(C’TOTAL: $’,3,7))

Figure 112. Sample IFTHEN Parameter

The report produced is:

DATE REG# ID
2008/08/16 0005 00002
SKU#: 013298 PRICE: $0000.69 QUANTITY: 004
SKU#: 510945 PRICE: $0017.03 QUANTITY: 001
TOTAL: $0019.79
DATE REG# ID
2008/08/17 0002 00003
SKU#: 212134 PRICE: $0003.49 QUANTITY: 003
TOTAL: $0010.47
DATE REG# ID
2008/08/17 0003 00001

2–206 Syncsort MFX Programmer’s Guide

 OUTREC

SKU#: 872567 PRICE: $00010.22 QUANTITY: 001

SKU#: 510945 PRICE: $0001.99 QUANTITY: 003
SKU#: 734018 PRICE: $0003.98 QUANTITY: 002
TOTAL: $0024.15
IFTHEN Processing Considerations
In an OUTFIL control statement, the IFTHEN parameter may be used with FTOV,
VLTRAIL, or VLTRIM. The IFTHEN parameter may not be used with CONVERT
or VTOF. Under the OUTFIL control statement, see “FTOV Parameter (Optional)”
on page 2-96 for a complete description of the FTOV parameter, “VLTRAIL
Parameter (Optional)” on page 2-96 for a complete description of the VLTRAIL
parameter, and “VLTRIM Parameter (Optional)” on page 2-97 for a complete
description of the VLTRIM parameter.
IFTHEN processing continues until:
• All IFTHEN parameters have been processed.
• A WHEN=(conditions) or WHEN=ANY subparameter condition is satisfied and
the HIT=NEXT subparameter is not included.
• Multiple output records are created with the / subparameter. Under the
OUTREC parameter of the OUTFIL control statement, see “[n]/” on page 2-94
for a complete description of the / subparameter.
The following is an example of the IFTHEN parameter:

OUTREC IFTHEN=(WHEN=INIT,
BUILD=(1,80,1,8,ZD,MUL,+107,DIV,+100,ZD)),
IFTHEN=(WHEN=(81,15,ZD,GT,+10000),
OVERLAY=(81:81,15,ZD,ADD,+0500,ZD),HIT=NEXT),
IFTHEN=(WHEN=(81,15,ZD,GT,+20000),
OVERLAY=(81:81,15,ZD,ADD,+2000,ZD),HIT=NEXT),
IFTHEN=(WHEN=ANY,
OVERLAY=(96:C’*’,97:81,15,ZD,MUL,+15,DIV,+100)),
IFTHEN=(WHEN=NONE,
OVERLAY=(97:81,15,ZD,MUL,+12,DIV,+100))

Figure 113. Sample IFTHEN Parameter

This OUTREC control statement refers to 80-byte input records containing a

salesperson’s weekly sales in dollars in the first field (1,8,ZD) of each record. There
are five IFTHEN parameters in the example and they reformat each input record
as follows:
• The first IFTHEN parameter uses WHEN=INIT to take the sales total in the
first field (1,8,ZD), increase it by 7%, and enter the result in a new 15-byte ZD
field in column 81.
• The second IFTHEN parameter uses WHEN=(conditions) to test if the newly
adjusted sales total in the field (81,15,ZD) is over $10,000. If it is, the second
IFTHEN parameter increases the total by $500 and replaces the result in that
field.

Syncsort MFX Programmer’s Guide 2–207

OUTREC

• The third IFTHEN parameter uses WHEN=(conditions) to test if the newly

adjusted sales total in the field (81,15,ZD) is over $20,000. If it is, the third
IFTHEN parameter increases the total by $2,000 and replaces the result in that
field.
• The fourth IFTHEN parameter uses WHEN=ANY to test the second and third
IFTHEN parameters. If one or both WHEN subparameter conditions are
satisfied, indicating that the salesperson is getting a bonus, the fourth IFTHEN
parameter inserts an “*” in column 96 to denote a bonus, calculates a
commission rate of 15%, and enters the result in column 97.
• The fifth IFTHEN parameter uses WHEN=NONE to test the second and third
IFTHEN parameters. If neither WHEN subparameter condition is satisfied,
indicating that the salesperson is not getting a bonus, the fifth IFTHEN
parameter calculates a commission rate of 12% and enters the result in column
97.

IFOUTLEN Parameter (Optional)

The IFOUTLEN parameter overrides the maximum record length, which is
automatically set by the IFTHEN parameter, and changes it to a specified value.
The IFOUTLEN parameter may only be used in conjunction with the IFTHEN
parameter.
The format of the IFOUTLEN parameter is illustrated below.

IFOUTLEN=n

Figure 114. IFOUTLEN Parameter Format

n The new maximum record length.

The IFOUTLEN parameter automatically makes the following changes to the
record to match the new length. A fixed-length or variable-length record longer
than n is truncated to n. In a fixed-length record shorter than n, the record is
padded with blanks to reach a length of n.

OVERLAY Parameter (Optional)

The OVERLAY parameter enables you to change particular columns and add fields
to the end of a record without rebuilding the entire record. When using the
OVERLAY parameter you only need to specify the columns you want to change.
The rest of the input record remains unchanged.
The format of the OVERLAY parameter is similar to that of the FIELDS parameter
of the INREC and OUTREC control statements. Under the OUTREC control
statement, see Figure 71 on page 2-136 for the format of the OVERLAY parameter.
The following exceptions apply:
• In the OVERLAY parameter the length l is always required, unlike one case of
the FIELDS parameter in which l is optional after p.

2–208 Syncsort MFX Programmer’s Guide

 OUTREC

• In the OVERLAY parameter for a variable-length record, the column value c: is

always required and must be set at 5 or greater, since c: is set to 1 by default
and positions 1 through 4 comprise the RDW. The RDW cannot be overlaid.
The OVERLAY parameter automatically makes the following changes to the output
record to accommodate any adjustments in length. In a variable-length record, the
RDW length is adjusted accordingly. In a fixed-length record, the record is padded
with blanks when necessary. Blanks also replace missing bytes in input fields.
Modifications to records will be made in the order of the OVERLAY parameters
specified. If you modify the same field more than once, the second and subsequent
modifications will apply to the previously modified field.
The following is an example of the OVERLAY parameter:

OUTREC OVERLAY=(9:9,4,PD,SUB,13,4,PD,PD,LENGTH=4,81:9,4,PD)

Figure 115. Sample OVERLAY parameter

This OUTREC control statement refers to an 80-byte record. The OVERLAY

parameter subtracts the Payments field (13,4,PD) from the Balance Due field
(9,4,PD). This updated Balance Due amount is entered in a new, displayable field
at the end of the record.
In an OUTFIL control statement, the OVERLAY parameter may be used with
FTOV, VLTRAIL, or VLTRIM. The OVERLAY parameter may not be used with
CONVERT or VTOF. Under the OUTFIL control statement, see “FTOV Parameter
(Optional)” on page 2-96 for a complete description of the FTOV parameter,
“VLTRAIL Parameter (Optional)” on page 2-96 for a complete description of the
VLTRAIL parameter, and “VLTRIM Parameter (Optional)” on page 2-97 for a
complete description of the VLTRIM parameter.

PARSE Parameter (Optional)

Use PARSE to extract variable-position and variable-length fields from records.
The resultant data will be placed into fixed-length parsed fields. The fixed-length
parsed fields are specified by %pp, where pp is an integer from 00 to 999. Therefore,
up to 1000 fixed-length parsed fields may be defined for each PARSE application.
The criteria for extracting variable fields are specified using the PARSE
subparameters. The resultant %pp fields may then be used to the same extent as
fixed fields, which have a fixed position p and a fixed-length l, in the FIELDS,
BUILD, or OVERLAY parameters associated with the statements.
For use of PARSE with IFTHEN and in the case of missing fields, see “PARSE with
IFTHEN” on page 2-217.

Syncsort MFX Programmer’s Guide 2–209

OUTREC

The syntax of PARSE is illustrated below:

 %pp= ( subparm 1 ,FIXLEN=l )   %pp= ( subparm 2 ,FIXLEN=l ) 

PARSE= (   ,  ... )
 %= ( subparm 1 [ ,FIXLEN=l ] )   %= ( subparm 2 [ ,FIXLEN=l ] ) 

subparm can be specified as follows:

 
  STARTAFT 1 =string   STARTAFT =string  
    2
 
  STARTAFT 1 =alphanum   STARTAFT 2 =alphanum  
  STARTAFT =BLANKS   STARTAFT =BLANKS  
 ABSPOS=p    1   2  
 ADDPOS=x   ,  STARTAT 1 =string  ,  STARTAT 2 =string  ... 
 SUBPOS=y    STARTAT =alphanum   STARTAT =alphanum  
  1   2  
  STARTAT 1 =BLANKS   STARTAT 2 =BLANKS 
  STARTAT =NONBLANK   STARTAT =NONBLANK  
   
  1  2

 
  ENDBEFR 1 =string   ENDBEFR 2 =string  
     
  ENDBEFR 1 =alphanum   ENDBEFR 2 =alphanum   
  ENDBEFR 1 =BLANKS   ENDBEFR 2 =BLANKS   , PAIR=APOST 
 ,  ,  ...   PAIR=QUOTE 
  ENDAT 1 =string   ENDAT 2 =string  
  ENDAT 1 =alphanum   ENDAT 2 =alphanum  
     [ ,REPEAT=m ]
  ENDAT 1 =BLANKS   ENDAT 2 =BLANKS  


Figure 116. PARSE and Subparameters

By default the first PARSE operation will begin at byte 1 for fixed-length records
and byte 5 for variable-length records. This represents the initial position of the
cursor within the record. The cursor can be repositioned to start the PARSE
operation through the use of the ABSPOS, ADDPOS, SUBPOS, STARTAFT, or
STARTAT subparameters. The PARSE operation to extract the field continues
until the ENDBEFR or ENDAT conditions are satisfied or, in their absence, for the
number of bytes specified in the FIXLEN subparameter. The cursor is advanced as
the result of processing the above subparameters.
A subsequent PARSE operation, by default, will begin at the byte where the cursor
was last positioned by the prior PARSE operation. This position can also be
modified as described above. Refer to the descriptions of the subparameters for
details on cursor position as a result of their operation.
Note: If TPF3 maintenance or PTF TY01443 has not been applied to Release 2.1,
then if the cursor advances past the end of the record, PARSE operation stops.
However, with PTF TY01443 or TPF3 applied, processing will revert to the method
used by Release 1.4, which continues processing the subsequent PARSE fields.

2–210 Syncsort MFX Programmer’s Guide

 OUTREC

The order in which the PARSE subparameters are processed is as follows:

• ABSPOS or ADDPOS or SUBPOS
• STARTAFT, STARTAT and PAIR
• ENDBEFR, ENDAT and PAIR
• FIXLEN

The following describes the PARSE subparameters:

%pp Defines the fixed-length parsed field with a unique
identifier pp, which is an integer from 0 to 999. A %pp
field can be defined only once in all PARSE subparam-
eters in an application. Therefore, up to 1000 unique
%pp fields can be defined and can be used more than
once in BUILD or OVERLAY. Note that the %pp fields
defined for a specific control statement can only be
used in the FIELDS, BUILD or OVERLAY parameter
for that statement.
Variables defined as %n are equivalent to %0n or
%00n (for example, %3 is equivalent to %03) and so
cannot both be defined in the same application.
% Specifies that the variable field will be ignored and not
extracted. The start position of the cursor for the next
parsed field is determined by the remaining subpa-
rameters.
ABSPOS=p Optionally specifies the absolute starting cursor posi-
tion p (bytes) for the parsed field. You can set p from 1
to 32752. You can use ABSPOS to override the starting
cursor position set by the previous parsed field; if it is
less than 5 for a variable-length record, then the cur-
sor position is defaulted to 5. (For fixed-length records,
the default position is at byte 1 for the first parsed
field. For variable-length records, the default position
is at byte 5 for the first parsed field.)
ADDPOS=x Optionally specifies that the start position of the cur-
sor will be at the current position plus x bytes added.
You can set x from 1 to 32752.
SUBPOS=y Optionally specifies that the start position of the cur-
sor will be at the current position minus x bytes sub-
tracted. You can set y from 1 to 32752. If the result is
less than 1 for a fixed-length record, then the cursor
position is set to 1. If the result is less than 5 for a
variable-length record, then the cursor position is set
to 5.

Syncsort MFX Programmer’s Guide 2–211

OUTREC

STARTAFT=string Optionally specifies a string, which indicates the start

of the parsed extraction of the variable field one byte
after the string (for example, a comma). The start posi-
tion of the cursor for the next parsed field is then set at
the byte after the string. If the string is not present,
then blank characters will be inserted into the current
parsed field and all subsequent parsed fields.
You can specify the string as a character string con-
stant (C'string') or hexadecimal string constant
(X'hh...hh'). For example, a comma would be specified
as STARTAFT=C','.
You can specify multiple instances and combinations of
any STARTAFT and STARTAT subparameter for a
single %pp parsed field. For example,
PARSE=(%01=(STARTAFT=C'/', STAR-
TAFT=C'<',STARTAT=C'*',FIXLEN=5)). From left to
right, the first STARTAFT or STARTAT criterion to be
satisfied will be the one to be implemented.
STARTAFT=alphanum Optionally specifies a set of alphanumeric characters,
which indicates the start of the parsed extraction of
the variable field one byte after a character from the
set has been found. The start position of the cursor for
the next parsed field is then set at the byte after the
character that was found. If no characters from the
specified set are present, then blank characters will be
inserted into the current parsed field and all subse-
quent parsed fields.
The choices for the alphanum character set include
lowercase, uppercase and numeric characters:
LC for lowercase characters a-z
UC for uppercase characters A-Z
MC for mixed case characters a-z and A-Z
LN for lowercase characters and numerics a-z and 0-9
UN for uppercase characters and numerics A-Z and
0-9
MN for mixed case characters and numerics a-z, A-Z
and 0-9
NUM for numerics 0-9
STARTAFT=BLANKS Optionally specifies the start of the parsed extraction
of the variable field at the first nonblank character
after one or more blanks. The start position of the cur-

2–212 Syncsort MFX Programmer’s Guide

 OUTREC

sor for the next parsed field is then set at the first non-
blank character. If a blank is not present, then blank
characters will be inserted into the current parsed
field and all subsequent parsed fields.
You can specify multiple instances and combinations of
any STARTAFT and STARTAT subparameter. See
“STARTAFT=string” above for further description.
STARTAT=string Optionally specifies a string, which indicates the start
of the parsed extraction of the variable field at the
position of, and including, the string. The start posi-
tion of the cursor for the next parsed field is then set at
the byte after the string. If the string is not present,
then blank characters will be inserted into the current
parsed field and all subsequent parsed fields.
You can specify the string as a character string con-
stant (C'string') or hexadecimal string constant
(X'hh...hh'). For example, a comma would be specified
as STARTAT=C','.
You can specify multiple instances and combinations of
any STARTAFT and STARTAT subparameter. See
“STARTAFT=string” above for further description.
STARTAT=alphanum Optionally specifies a set of alphanumeric characters,
which indicates the start of the parsed extraction of
the variable field at the position of, and including, the
character from the set that was found. The start posi-
tion of the cursor for the next parsed field is then set at
the byte after the character that was found. If no char-
acters from the specified set are present, then blank
characters will be inserted into the current parsed
field and all subsequent parsed fields.
See STARTAFT=alphanum for a description of the
choices for the alphanum character set.
STARTAT=BLANKS Optionally specifies the start of the parsed extraction
of the variable field at the position of, and including,
the first blank character. The start position of the cur-
sor for the next parsed field is then set at the first non-
blank character. If a blank is not present, then blank
characters will be inserted into the current parsed
field and all subsequent parsed fields.
You can specify multiple instances and combinations of
any STARTAFT and STARTAT subparameter. See
“STARTAFT=string” above for further description.

Syncsort MFX Programmer’s Guide 2–213

OUTREC

STARTAT=NONBLANK Optionally specifies the start of the parsed extraction

of the variable field at the position of, and including,
the first nonblank character. The start position of the
cursor for the next parsed field is then set at the first
nonblank character. If a nonblank is not present, then
blank characters will be inserted into the current
parsed field and all subsequent parsed fields.
You can specify multiple instances and combinations of
any STARTAFT and STARTAT subparameter. See
“STARTAFT=string” above for further description.
ENDBEFR=string Optionally specifies a string, which indicates the end
of the parsed extraction of the variable field one byte
before the string (for example, a comma). The start
position of the cursor for the next parsed field is then
set at the byte after the string.
If the string is not present, then data from the field
will continue to be extracted up until the end of the
record. Blank characters will be inserted into all sub-
sequent parsed fields.
You can specify the string as a character string con-
stant (C'string') or hexadecimal string constant
(X'hh...hh'). For example, a comma would be specified
as ENDBEFR=C','.
You can specify multiple instances and combinations of
any ENDBEFR and ENDAT subparameter for a single
%pp parsed field. For example, PARSE=(%01=(END-
BEFR=C'/', ENDBEFR=C'<',ENDAT=C'*',FIX-
LEN=5)). From left to right, the first ENDBEFR or
ENDAT criterion to be satisfied will be the one to be
implemented.
ENDBEFR=alphanum Optionally specifies a set of alphanumeric characters,
which indicates the end of the parsed extraction of the
variable field one byte before a character from the set
has been found. The start position of the cursor for the
next parsed field is then set at the byte after the char-
acter that was found. If no characters from the speci-
fied set are present, then data from the field will
continue to be extracted up until the end of the record.
Blank characters will be inserted into all subsequent
parsed fields.
See STARTAFT=alphanum for a description of the
choices for the alphanum character set.

2–214 Syncsort MFX Programmer’s Guide

 OUTREC

ENDBEFR=BLANKS Optionally specifies the end of the parsed extraction of

the variable field one byte before a blank character is
encountered. The start position of the cursor for the
next parsed field is then set at the first nonblank char-
acter after the blank (or group of blanks).
If a blank character is not present, then data from the
field will continue to be extracted up until the end of
the record. Blank characters will be inserted into all
subsequent parsed fields.
You can specify multiple instances and combinations of
any ENDBEFR and ENDAT subparameter. See “END-
BEFR=string” above for further description.
ENDAT=string Optionally specifies the end of the parsed extraction of
the variable field at the position of, and including, the
last string character. The start position of the cursor
for the next parsed field is then set at the byte after
the string.
If the string is not present, then data from the field
will continue to be extracted up until the end of the
record. Blank characters will be inserted into all sub-
sequent parsed fields.
You can specify the string as a character string con-
stant (C'string') or hexadecimal string constant
(X'hh...hh'). For example, a comma would be specified
as ENDAT=C','.
You can specify multiple instances and combinations of
any ENDBEFR and ENDAT subparameter. See “END-
BEFR=string” above for further description.
ENDAT=alphanum Optionally specifies the end of the parsed extraction of
the variable field at the position of, and including, a
character from the specified alphanum set that was
found. The start position of the cursor for the next
parsed field is then set at the byte after the character
that was found. If no characters from the specified set
are present, then data from the field will continue to
be extracted up until the end of the record. Blank char-
acters will be inserted into all subsequent parsed
fields.
See STARTAFT=alphanum for a description of the
choices for the alphanum character set.
ENDAT=BLANKS Optionally specifies the end of the parsed extraction of
the variable field at the position of, and including, the

Syncsort MFX Programmer’s Guide 2–215

OUTREC

last blank character. The start position of the cursor

for the next parsed field is then set at the first non-
blank character after the blank (or group of blanks).
If a blank character is not present, then data from the
field will continue to be extracted up until the end of
the record. Blank characters will be inserted into all
subsequent parsed fields.
You can specify multiple instances and combinations of
any ENDBEFR and ENDAT subparameter. See “END-
BEFR=string” above for further description.
PAIR=APOST Optionally specifies that all characters between pairs
of apostrophes ('characters') be ignored when search-
ing for a string or blanks. If only one apostrophe is
present, all characters to the right of the apostrophe
will be ignored.
PAIR=QUOTE Optionally specifies that all characters between pairs
of quotes ("characters") be ignored when searching for
a string or blanks. If only one quote is present, all
characters to the right of the quote will be ignored.
FIXLEN=l Specifies the length l (1 to 32752) in bytes of the %pp
field. FIXLEN is required when used with %pp, but
optional when used with %. If ENDBEFR or ENDAT is
not specified, then FIXLEN indicates the end of the
parsed extraction of the variable field at the end of
length l. Thus, the start position of the cursor for the
next parsed field is set at the next byte following the
length.
If the PARSE operation produces a field less than FIX-
LEN, the parsed field will be left-justified and padded
on the right with the difference in blank characters. If
the length of the parsed field is greater than l, the data
will be truncated after l bytes.
REPEAT=m Optionally is a shorthand way to specify the repetition
of the currently defined parsed field. m can be from 2
to 1000 and defines the total number of identical
parsed fields. When used with a % field, m consecutive
fields that will be ignored are created. When used with
a %nn field, consecutively numbered parsed fields
from nn to nn+m-1 will be defined. For instance, speci-
fying %3=(your_subparms,REPEAT=4) will create 4
identically defined parsed fields:
%3=(your_subparms),%4=(your_sub-
parms),%5=(your_subparms),%6=(your_subparms)

2–216 Syncsort MFX Programmer’s Guide

 OUTREC

When defining a %nn field with REPEAT, be sure that

you have not defined duplicate %nn fields elsewhere in
your control statements. For the above %3 example,
you may not define %4, %5 or %6 anywhere else. You
must also ensure that the maximum %nn number cre-
ated by REPEAT is %999.
PARSE with IFTHEN
You can use %pp parsed fields in IFTHEN expressions. If the %pp field is defined in
a WHEN=INIT, WHEN=(conditions), WHEN=ANY, or WHEN=NONE expression,
it can be used in the IFTHEN BUILD or IFTHEN OVERLAY of that expression.
Additionally, for WHEN=INIT, the %pp fields can be used in any subsequent
IFTHEN expression. See “IFTHEN Parameter (Optional)” on page 2-198 for
further description of the IFTHEN parameter.
A sample application of using PARSE with IFTHEN is when the parse cursor needs
to be reset to the default position at the beginning of the record, as in the case of
variable records with missing fields. For each WHEN=INIT statement
implemented with PARSE, the cursor position is set to byte 1 for fixed-length
records and byte 5 for variable-length records. Using PARSE without IFTHEN, a
search resulting in a missing field would cause any subsequent fields to be
overlooked and not properly parsed into %pp fields. However, using IFTHEN
PARSE, each search would reset the cursor to the beginning of the record and fields
could be properly parsed into %pp fields independent of each other.

Sample Statements Using PARSE

Example 1: Stock Portfolio
A file with comma-delimited records for a stock portfolio contains fields for stock
symbol, current price, and today’s change amount:

DIS,34.56,+1.09
T,37.05,-.42
GOOG,449.12,-11.62

Syncsort MFX Programmer’s Guide 2–217

OUTREC

To format this information into fixed-length columns so that the data can be
properly sorted and displayed, the following INREC and SORT statements may be
used:

INREC PARSE=(%1=(ENDBEFR=C',',FIXLEN=4), * STOCK SYMBOL (MAX LEN 4)

%2=(ENDBEFR=C',',FIXLEN=6), * CURRENT PRICE (MAX LEN 6)
%3=(FIXLEN=1), * SIGN OF TODAY'S CHANGE
%4=(ENDBEFR=C' ',FIXLEN=5)), * CHANGE AMOUNT (MAX LEN 5)
BUILD=(01:%1, * STOCK SYMBOL
07:%2,JFY=(SHIFT=RIGHT), * CURRENT PRICE
15:%3, * SIGN OF TODAY'S CHANGE
16:%4,JFY=(SHIFT=RIGHT)) * CHANGE AMOUNT
SORT FIELDS=(1,4,CH,A) * SORT BY STOCK SYMBOL

Figure 117. Example 1, INREC Statement with PARSE

The ENDBEFR subparameters for the %1 and %2 parsed fields capture the data in
the first two fields in the input records up until the comma delimiters and
reposition the cursor after the commas, while ENDBEFR for %4 works similarly for
the last field in each record. FIXLEN sets the maximum output length for each
field. %3 is used to strip the sign off the change amount, so that the numeric part of
the amount can be right-justified. The BUILD parameter is used to right-justify the
numeric data into columns and to add spacing between the numbers. Using INREC
allows the data to be sorted by stock symbol, producing the following output:

DIS 34.56 + 1.09

GOOG 449.12 -11.62
T 37.05 - .42

Example 2: Name and Address Data

A file has records with name and address information in a keyword format

NAME1=GEORGE;NAME2=BUSH;ADDR1=OVAL OFFICE;ADDR2=1600 PENNSYLVANIA

AVE;CITY=WASH
NAME1=WILLIAM;MI=J;NAME2=CLINTON;ADDR1=15 OLD HOUSE LN;CITY=CHAPPAQUA;STATE=NY
NAME1=GEORGE;MI=H;NAME2=BUSH;CITY=HOUSTON;STATE=TX

PARSE may be used to search for each keyword and extract the data into fixed-
length fields in a reconstructed record. In this example, some of the keywords in
certain records may be missing. This normally would cause the cursor to be moved
to the end of the record, so that the search for the next keyword fails. But, by using
PARSE with an IFTHEN WHEN=INIT separately for each field, this problem can

2–218 Syncsort MFX Programmer’s Guide

 OUTREC

be avoided because the cursor is reset to the beginning of the record for each new
PARSE.

INREC IFTHEN=(WHEN=INIT, * USE WHEN=INIT ONCE FOR EACH KEYWORD IN DATA

PARSE=(%1=(STARTAFT=C'NAME1=',ENDBEFR=C';',FIXLEN=12))),
IFTHEN=(WHEN=INIT,
PARSE=(%2=(STARTAFT=C'MI=',ENDBEFR=C';',FIXLEN=1))),
IFTHEN=(WHEN=INIT,
PARSE=(%3=(STARTAFT=C'NAME2=',ENDBEFR=C';',FIXLEN=12))),
IFTHEN=(WHEN=INIT,
PARSE=(%4=(STARTAFT=C'ADDR1=',ENDBEFR=C';',FIXLEN=24))),
IFTHEN=(WHEN=INIT,
PARSE=(%5=(STARTAFT=C'ADDR2=',ENDBEFR=C';',FIXLEN=24))),
IFTHEN=(WHEN=INIT,
PARSE=(%6=(STARTAFT=C'CITY=',ENDBEFR=C';',FIXLEN=12))),
IFTHEN=(WHEN=INIT,
PARSE=(%7=(STARTAFT=C'STATE=',ENDBEFR=C';',FIXLEN=2))),
* AFTER EXTRACTING THE DATA FOR EACH KEYWORD,
* ARRANGE IT IN FIXED COLUMNS
IFTHEN=(WHEN=NONE,
BUILD=(1:%1,14:%2,16:%3,29:%4,54:%5,79:%6,92:%7))
SORT FIELDS=(92,2,CH,A) * SORT BY "STATE"

Figure 118. Example 2, INREC Statement with IFTHEN PARSE

This produces the following output, where blanks are used for each missing field:

GEORGE BUSH OVAL OFFICE 1600 PENNSYLVANIA AVE WASH

WILLIAM J CLINTON 15 OLD HOUSE LN CHAPPAQUA NY
GEORGE H BUSH HOUSTON TX

Sample OUTREC Control Statements

Example 1
The following example illustrates how the OUTREC control statement can be used
to insert binary zeros and blanks into the record.

OUTREC FIELDS=(1:4Z,5:20,10,23:44,28,10X)

Figure 119. Example 1, Sample OUTREC Control Statement

This OUTREC control statement defines a 60-byte record as follows:

• Four binary zeros are inserted in the first 4 bytes of the record (4Z).
• The next field begins in position 5. This field began in position 20 before
OUTREC processing and is 10 bytes long (5:20,10).
• Eight blanks are inserted before the next field, which is positioned at byte 23.
MFX automatically inserts blanks in the unused positions between fields.

Syncsort MFX Programmer’s Guide 2–219

OUTREC

• The next field begins in position 23. This field began in position 44 before
OUTREC processing and is 28 bytes long (23:44,28).
• Ten blanks are inserted in the last 10 bytes of the record (10X).
Example 2
The following example illustrates how the OUTREC control statement can be used
to convert and edit numeric fields.

OUTREC FIELDS=(1,50,64,4,PD,M2,68,6,ZD,
EDIT=($I,IIT.TTS),SIGNS=(,,+,-))

Figure 120. Example 2, Sample OUTREC Control Statement

This OUTREC control statement defines a 70-byte output record as follows:

• The first field (1,50) begins in position 1. This field began in position 1 before
OUTREC processing and is 50 bytes long.
• The next field (64,4) begins in position 51. This packed decimal field began in
position 64 before OUTREC processing and is 4 bytes long. After being
converted and edited by editing mask M2 (64,4,PD,M2) the resulting field will
be 10 bytes long. However, the number of digits that will actually print will
depend on the number of leading zeros, if any, because this mask specifies that
only three digits must print whether or not they are leading zeros. Moreover,
this mask specifies that a minus sign print after the number if it is negative
and a blank print after the number if it is positive.
• The last field (68,6) begins in position 61. This zoned decimal field began in
position 68 before OUTREC processing and is 6 bytes long. The EDIT and
SIGNS subparameters (EDIT=($I,IIT.TTS),SIGNS=(,,+,-)) specify a 10-byte
field because 4 additional bytes are needed for the dollar sign, the comma, the
decimal point and the trailing plus or minus sign. Note that if the first three
digits are leading zeros, they will be suppressed.
Example 3
This example uses the OUTREC control statement to convert numeric data from
one format to another.

OUTREC FIELDS=(1,10,ZD,PD,
11,4,FI,ZD,LENGTH=8)

Figure 121. Example 3, Sample OUTREC Control Statement

This OUTREC control statement defines a 14-byte output record as follows:

• The first field (1,10,ZD,PD) begins in position 1. This field was a 10-byte ZD
field that began in position 1 before OUTREC processing. It will be converted to

2–220 Syncsort MFX Programmer’s Guide

 OUTREC

a 6-byte PD field in the output record, because 6 bytes are required to contain
10 decimal digits as a PD field.
• The next field (11,4,FI,ZD) begins in position 7. This field was a 4-byte FI field
that began in position 11 before OUTREC processing. It will be converted to an
8-byte ZD field in the output record. Normally 10 ZD bytes would be required to
contain the 10 decimal digits that may be represented by a 4-byte FI field, but
the LENGTH=8 parameter overrode the output length. If there are more than 8
decimal digits in any of the 11,4,FI fields, those digits will be truncated on the
left in the output record.
Note that ZD output is not the same as printable output using editing masks.
High order zeros will appear as zeros in a ZD field, while they appear as blanks
when using the default M0 mask, as well as most other masks. The sign
indicator in a ZD field is placed in the first 4 bits of the rightmost byte, and not
as a separate printable sign.
Example 4
This OUTREC example uses arithmetic and function operators to do algebraic
calculations.
New 8-byte PD fields are required in each record containing the maximum and
average of fields A, B, and C. Another new 5-byte printable field is required
containing field D as a percentage of field E. The field definitions are:

Field A: 1,4,PD

Field B: 5,8,ZD

Field C: 13,4,FI

Field D: 25,4,PD

Field E: 29,4,PD

The OUTREC control statement to accomplish this would be:

Syncsort MFX Programmer’s Guide 2–221

OUTREC

OUTREC FIELDS=(1,36, Retain existing fields

40:(01,4,PD,ADD, Field A plus
05,8,ZD,ADD, Field B plus
13,4,FI), Field C
DIV,+3, divide by 3 to get average
PD, output as 8-byte PD field
*
50:01,4,PD,MAX, Determine maximum of Field A and
05,8,ZD,MAX, Field B and
13,4,FI, Field C
PD, output as 8-byte PD field
* 60:+100,MUL, 100 times
25,4,PD,DIV, Field D divided by
29,4,PD, Field E
LENGTH=5) output as printable 5-byte field
* using default M0 mask

Figure 122. Example 4, Sample OUTREC Control Statement

This OUTREC control statement defines a 64-byte output record as follows:

• The first field (1,36) retains the complete contents of the input record.
• The second output field begins in position 40. An arithmetic calculation is done
using three different numeric input fields and the constant +3 to compute the
arithmetic average. This is an expression that is considered to contain 15
decimal digits. The output is requested as a PD field. The length of this field
will be 8 bytes, since that is the length required to contain 15 decimal digits.
• The third output field begins in position 50. Multiplying numeric Field D by 100
before dividing by numeric Field E gives the desired percentage number, which
is considered to contain 15 decimal digits. No output format or editing mask is
specified, so the default mask M0 is used to create printable output.
LENGTH=5 is specified to reduce the default length of the output field from 16
to 5, since it is known that the percentage number will not be large.
Example 5
This OUTREC control statement uses DT1, TM1, and edit masks to convert SMF
date and time values to appropriate formats.

OUTREC FIELDS=(1,4,DT1,EDIT=(TTTT/TT/TT),

3X,5,4,TM1,EDIT=(TT:TT:TT))

Figure 123. Sample OUTREC Control Statement

The following shows how the output would be formatted:

2002/07/04 07:22:12
2002/07/04 05:15:25

2–222 Syncsort MFX Programmer’s Guide

 OUTREC

2002/07/05 11:37:39
2002/07/05 16:42:28
Example 6
This OUTREC control statement illustrates the use of the &DATE1(c) and
&TIME1(c) parameters in an MFX run on June 9, 2002 at 04:16:29 p.m.

OUTREC FIELDS=(8,20,24:&DATE1(' '),X,&TIME1(:))

Figure 124. Sample OUTREC Control Statement

The output would include data from the input record in the first twenty columns
followed by the run-time date and time starting in column 24. The date and time
would appear as '2002 06 09 16:16:29'.
Example 7
The following control statements illustrate two of the options of the TRAN
subparameter.
This OUTREC control statement uses TRAN=LTOU to translate the letters in
positions 1-5 of each output record from lowercase to uppercase.

OUTREC FIELDS=(1,5,TRAN=LTOU)

Figure 125. Sample OUTREC Control Statement

For example, 'Ab,Cd' would translate to 'AB,CD'.

This OUTREC control statement uses TRAN=ALTSEQ to translate each binary
zero (X'00') in columns 1-5 to an asterisk (X'5C') in positions 1-5.

ALTSEQ CODE=(005C)

OUTREC FIELDS=(1,5,TRAN=ALTSEQ)

Figure 126. Sample OUTREC Control Statement

Comprehensive examples illustrating the OUTREC control statement and the

OUTREC parameter of the OUTFIL control statement are provided in “Chapter 3,
How to Use the MFX Data Utility Features”.

Sample OUTREC Control Statements with CENTWIN Processing

For century window processing, data conversion is determined by the century
window defined by the CENTWIN parameter.
The following provides examples of data conversion with CENTWIN:
Example 1

Syncsort MFX Programmer’s Guide 2–223

OUTREC

A 2-digit year field in character format at position 20 in the input record could be
expanded with the following specification:

OUTREC FIELDS=(1,19,* Copies first 19 bytes of record

20,2,Y2C,* Converts 2-digit year data to 4-digit year
22,59)* Copies remaining 59 bytes

Figure 127. Example 1, OUTREC Control Statement with Year Data

Note that the expansion of the year data from 2 to 4 digits increases the output
record length by 2 bytes compared to the input record length.
The CENTWIN setting determines the century of the 2-digit year field. If
CENTWIN=1980, then a year field in the input record would be converted as
follows:

SORTIN Input OUTREC Output

13 2013

79 2079

80 1980

92 1992

Example 2
Consider the following packed decimal date field at position 20 in the input record:
yymmdd = X'0yymmddC'
Suppose you want to output a displayable 4-digit year in character format in the
form
mm/dd/yyyy
To accomplish this, specify the following OUTREC control statement:

OUTREC FIELDS=(1,19, * Copies first portion of record

21,2,PD0,M11, * Converts X'ymmd' to X'mm' then C'mm'
C'/', * Inserts slash
22,2,PD0,M11, * Converts X'mddC' to X'dd'then C'dd'
C'/', * Inserts slash
20,2,Y2P, * Converts X'0yym' to X'yy' then C'yyyy'
24,76) * Copies rest of record

Figure 128. Example 2, OUTREC Control Statement with Year Data

The 4-digit year output from the input year field (20,2,Y2P) depends on the
CENTWIN setting. The following sample of input and output data shows the case
for CENTWIN=1980:

2–224 Syncsort MFX Programmer’s Guide

 OUTREC

SORTIN Input Date OUTREC Output Date Field

Field

X'0800329C' 03/29/1980

X'0790603C' 06/03/2079

Example 3
To expand a 3-byte packed decimal date field of the form X'yyddds', at position 20 in
the input record, to a 4-byte packed field of the form X'yyyyddds' that contains a
prefixed century value, specify an OUTREC control statement such as the
following:

OUTREC FIELDS=(1,19, * Copies first portion of record

20,1,Y2ID, * Converts X'yy' to X'yyyy'
21,60) * Copies rest of record starting with
* the X'ddds' of the date field

Figure 129. Example 3, OUTREC Control Statement with Year Data

Note that in the above example the output record length will be 1 byte larger than
the input record length. The following sample of input and output data shows the
effect for CENTWIN=1980:

SORTIN Input Date OUTREC Output Date Field

Field

X'79' X'2079'

X'80' X'1980'

Example 4
To expand a 4-byte packed decimal date field of the form X'0yymmdds', at position
20 in the input record, to a 5-byte field of the form X'0yyyymmdds' that contains a
prefixed century value, specify an OUTREC control statement such as the
following:

OUTREC FIELDS=(1,19, * Copies first portion of record

20,2,Y2IP, * Converts X'0yym' to X'0yyyym'
22,59) * Copies rest of record starting with
* * the X'mdds' of the date field

Figure 130. Example 4, OUTREC Control Statement with Year Data

Syncsort MFX Programmer’s Guide 2–225

OUTREC

As with Y2ID conversion, the output record length will be 1 byte larger than the
input length. The following sample of input and output data shows the effect for
CENTWIN=1980:

SORTIN Input Date OUTREC Output Date Field

Field

X'0790' X'020790'

X'0801' X'019801'

Example 5
Consider a 2-byte character or zoned decimal field that may contain either valid
numeric year data or characters that identify the record as a header or trailer.
Header records in the example are identified by zeros (X'00') or a blank (X'40') in
the first byte of the year field, while trailer records are identified by binary ones
(X'FF') in the first byte of the field. The Y2S format will treat the valid year data
normally, in the same way as the Y2C or Y2Z formats would treat the data, but the
year fields of header and trailer records will be converted to a 4-digit form padded
on the left with data identical to the data in the first byte of the input field.
Typically this type of conversion is needed when a Y2S SORT or MERGE field is
used to collate the records so that header/trailer records in the output remain at
the start or end of the file. An OUTREC control statement such as the following
could be used.

OUTREC FIELDS=(1,19, * Copies first portion of record

20,2,Y2S, * Converts C'yy' to C'yyyy' and pads
* fields that identify header/trailer records
22,59) * Copies the remaining fields

Figure 131. Example 5, OUTREC Control Statement with Year Data

As with Y2C or Y2Z, the output record length will be 2 bytes larger than the input
record length.
For CENTWIN=1990, the sorted Y2S field would be converted as follows:

SORTIN Input Date Field OUTREC Output Date Field

X'4001' X'00000000' (from 4th input record)

X'F9F8' X'40404001' (from 1st input record)

X'F0F3' X'F1F9F9F8' (from 2nd input record)

X'0000' X'F2F0F0F3' (from 3rd input record)

X'FFFF' X'FFFFFFFF' (from 5th input record)

2–226 Syncsort MFX Programmer’s Guide

 RECORD

RECORD Control Statement

The RECORD control statement provides record length and format information. It
is required in the following situations:
• MFX is invoked by a program using an in-memory E15 or E32 exit routine.
• An E15 or E35 exit routine changes the record length.

RECORD Control Statement Format

The format of the RECORD control statement is illustrated below:

F 
RECORD TYPE=   [,LENGTH=(l 1 ,l 2 ,l 3 ,l 4 ,l 5 ,l 6 ,l 7 )]
V 

Figure 132. RECORD Control Statement Format

TYPE Parameter (Optional)

The TYPE parameter can be used to indicate the record format. TYPE=F indicates
fixed-length records; TYPE=V indicates variable-length records. TYPE=FB or
TYPE=VB can be specified but the 'B' is ignored.
TYPE should be specified if SORTIN is VSAM. If TYPE is not provided, the
SORTOUT RECFM will be examined to determine the SORTIN TYPE. If no
SORTOUT RECFM is found, TYPE=V will be assumed if SORTOUT is VSAM and
TYPE=F if there is no SORTOUT or SORTOUT is non-VSAM.
Note: If the TYPE specification differs from the RECFM DCB parameter for the
SORTIN/SORTINnn DD statement, the latter takes precedence.

LENGTH Parameter (Conditionally Required)

The LENGTH parameter, usually optional, is required whenever the RECORD
control statement is required.
The LENGTH parameter specifies the length of the record at various points during
the processing of the application.
The number of length values can vary from 1 to 7. Only the l1, l2 and l3 values
should be specified for fixed-length records and for merge or copy applications. All
seven length values can be specified for variable-length sorts. If l1 is the only value
specified, parentheses are optional. If l1 and additional length values are specified,
they all must be enclosed in parentheses.
The length values are positionally dependent. An extra comma must indicate a
missing length value between any two that are specified. Commas need not follow

Syncsort MFX Programmer’s Guide 2–227

RECORD

the final length value specified. For example, if LENGTH=(l1,,,l4) is specified, the
omitted values are understood to be l2 and l3.
The l1,...,l7 variables specify the following:
l1 The maximum record input length of the logical records. For vari-
able-length records, this is the length of the longest logical record
plus the 4-byte Record Descriptor Word. The 4-byte RDW must be
included, even if the input is a VSAM file. The maximum record
length cannot exceed 32,760 for fixed-length records and 32,767 for
variable-length records. An LRECL value specified on the SOR-
TIN/SORTINnn DD statement or the data set label will override the
l1 value for fixed-length records. For variable-length records, the
higher value (LRECL or l1) is used. This is ignored in a join applica-
tion.
l2 The maximum length of the logical records after E15 processing. An
omitted l2 value defaults to the l1 value and indicates that the maxi-
mum record length has not been changed by an E15 exit. If there is
no E15 exit, an l2 value which is smaller than the l1 value or the
LRECL specified on the SORTIN/SORTINnn DD statement or data
set label will truncate the records. This truncation will occur after
the record is read from SORTIN. This is ignored in a join application.
l3 The maximum length of the logical records after E35 processing. If
the l3 value is omitted, the default is either the l2 value, or, if an
INREC and/or OUTREC control statement is specified, the record
length after INREC/OUTREC processing. Note that it is not neces-
sary to specify an l3 value to reflect a length change due to INREC or
OUTREC processing; the revised record length is calculated auto-
matically. However, it is necessary to specify an l3 value if exit E35
has altered the record length.
The LRECL value specified in the SORTOUT DD statement should
either correspond to the l3 value or the LRECL specification should
be omitted. In the latter case, MFX will automatically calculate the
correct LRECL value.
The l3 value is ignored if there is no E35 exit, so it is not possible to
use the l3 value to truncate or pad the records.
l4 The minimum length of the variable-length logical records plus the
4-byte Record Descriptor Word. An omitted l4 value defaults to the
length from the beginning of the record to the end of the last field ref-
erenced by any control statement.
l5 The most frequent record length of the variable-length records. Spec-
ify this length value to optimize the size of the segment, i.e., the
fixed-length block of main storage, used to contain variable-length
records.

2–228 Syncsort MFX Programmer’s Guide

 RECORD

l6 The average work space required by each record, as reported by the

HISTOGRM utility program.
l7 The segment length recommended by the HISTOGRM utility pro-
gram. If l7 is omitted, the SIZE parameter on the SORT control state-
ment may be used to determine the impact of segment size on sort
performance. Assuming the SIZE parameter reports a SORTIN data
set of at least 10,000 records, MFX may sample the first 100-200
records to calculate an approximate segment size. An installation
may decide to allow record sampling for smaller files.

Rules for Specifying the Length Parameter

Observe the following rules when specifying length values:
• All length values for variable-length records must include 4 bytes for the
Record Descriptor Word.
• The l1, l2, and l3 values must represent the maximum record lengths and the l4
value must represent the minimum record length. If MFX encounters a record
which exceeds the maximum length or is shorter than the minimum length, the
application will either terminate abnormally or produce unpredictable results.

Sample RECORD Control Statements

RECORD TYPE=F,LENGTH=(80,,60)

Figure 133. Sample RECORD Control Statement

This sample RECORD control statement defines the record as follows:

• The file contains fixed-length records.
• The input record length (l1) is 80 bytes.
• A comma represents the omitted l2 value because an E15 exit does not change
the record length.
• The record length after INREC/OUTREC and/or E35 processing is 60 bytes.
The SORTOUT LRECL should either be specified as 60 or omitted. If it is
omitted, MFX will automatically supply the correct value.

RECORD TYPE=V,LENGTH=(400,300,250,l20,200,280,230)

Figure 134. Sample RECORD Control Statement

This sample RECORD control statement defines the record as follows:

• The file contains variable-length records. All length values include 4 bytes for
the Record Descriptor Word.

Syncsort MFX Programmer’s Guide 2–229

RECORD

• The maximum input record length is 400 bytes.

• The maximum record length after E15 processing is 300 bytes.
• The maximum record length after INREC/OUTREC and/or E35 processing is
250 bytes.
• The minimum record length is 120 bytes.
• The most frequent record length is 200 bytes.
• The average work space required for each record is 280 bytes, as reported by the
HISTOGRM utility program.
• The segment length recommended by HISTOGRM is 230 bytes.
In the above example, the l4, l5, l6 and l7 values will be ignored if the application is a
merge or copy.

2–230 Syncsort MFX Programmer’s Guide

 REFORMAT

REFORMAT Control Statement

The REFORMAT control statement defines the record layout to be produced by the
join processing specified on an application’s JOINKEYS control statements.
Use the REFORMAT control statement to specify which fields from the SORTJNF1
and SORTJNF2 files are to be included in each record created by the join operation.
The REFORMAT control statement is normally required if JOINKEYS is specified.
It is optional if a JOIN control statement with the ONLY option has been specified
in a join application since no records will actually be joined. In that instance, if a
REFORMAT control statement is not provided and ONLY the unpaired records
from one join input (SORTJNF1 DD or SORTJNF2 DD) are requested on the JOIN
control statement, the records will not be reformatted and the record type and
length of the join input file will be retained.
If a REFORMAT control statement is not provided and ONLY the unpaired records
from both join inputs are requested, the resultant records will be variable-length,
regardless of the record formats of the join input data sets, and the record length
will be the maximum of any fixed-length input file record length plus four (for an
RDW) and any variable-length input file record length.

REFORMAT Control Statement Format

The format of the REFORMAT control statement is illustrated below:

REFORMAT FIELDS= ( Fn:p 1 ,l 1 [ ,[Fn:]p 2 ,l 2 ] [ ,? ] [ ,[Fn:]p 3 ,l 3 ]... [ ,[Fn:]p m ] [ ,Fn:p n ] ) [ ,FILL=f ]

Figure 135. REFORMAT Control Statement Format

FIELDS Parameter (Required)

The FIELDS parameter specifies fields to be included in the record produced by the
join function.
Each data field specified in the FIELDS parameter is identified by the file it
originates from Fn, its position p and length l.
Fn: The Fn value indicates the input file from which the data field should be
copied. Code ‘F1’ for SORTJNF1 and ‘F2’ for SORTJNF2. This field is
optional after the first field specification. By default, the file of the prior
field specification will be used to determine the current field specifica-
tion.
If your join application requests only the unpaired records from one join
input through the JOIN statement, then you may not reference the
other join input file in the FIELDS parameter, since no records from that
file will ever be selected. For example, if “JOIN UNPAIRED,F1,ONLY” is
specified, then F2 may not be used in the FIELDS parameter.

Syncsort MFX Programmer’s Guide 2–231

REFORMAT

p The position value indicates the first byte of the field relative to the
beginning of the input record.
l The length value indicates the length of the field.
? This symbol is used to place a one-byte indicator in the reformatted
record that indicates whether the reformatted record is a paired or an
unpaired joined record. The indicator will be set to one of three different
printable values:

“B” if the reformatted record is a paired record

“1” if the reformatted record is an unpaired record created from the F1
file
“2” if the reformatted record is an unpaired record created from the F2
file

? may only be used once in the FIELDS parameter. If it is followed by

any p,l fields, you must specify the Fn: subparameter. If a variable-
length reformatted record is created , ? must be placed before any field
specifying the variable part of the record.

Specifying the FIELDS Parameter for Variable-Length Records

If the REFORMAT statement only defines p,l fields and/or a ? field, then the output
of the join will be a fixed-length record. If a variable-length record format is desired
when one or both input files are variable-length, then the first p,l REFORMAT field
must be 1,4 (from either SORTJNFn input file) to define the RDW. This p,l
specification of 1,4 must reference an Fn that is a variable-length file. The variable
portion of the record must then be specified as the last REFORMAT field by coding
a position p without a length l. If both files are variable-length, then the variable
portion from each of the variable-length input files may be specified once at the end
of the REFORMAT statement.

REFORMAT FIELDS=(F1:1,4,F1:10,10,F2:25,3,?,F1:40,F2:50)

Figure 136. Sample REFORMAT Statement

FILL Parameter (Optional)

The FILL parameter defines a fill byte to be used for any missing p,l field bytes.
The format of the FILL parameter is illustrated below (f specifies the fill byte):

FILL=f

Figure 137. FILL Format

2–232 Syncsort MFX Programmer’s Guide

 REFORMAT

f can be specified as either a character or hexadecimal value. Specify either C'x'

where x is a single EBCDIC character or X'hh' where hh represents a hexadecimal
digit pair (00-FF).
The need for a fill byte can arise from two conditions:
• A portion or an entire p,l field specification is missing due to a short variable-
length record.
• A JOIN UNPAIRED was used and the REFORMAT FIELDS specification
requires a field from the file that is not being used to generate the current
joined record.
The default FILL character is a blank. Binary zeros will be used instead of the
FILL character for the first four bytes of a variable-length record requiring FILL
processing. This indicates that a record was not present for the REFORMAT due to
JOIN UNPAIRED.

Sample REFORMAT Control Statement

JOIN UNPAIRED,F1
REFORMAT FIELDS=(F1:10,10,F2:12,5),FILL=C'0'

Figure 138. Sample JOIN and REFORMAT Statements

In this example, if a record is found in SORTJNF1 that does not match a record in
SORTJNF2, the unpaired record will be included in the join output and would look
as follows:

Position Value

1-10 Contents of the SORTJNF1 record positions 10 through 19.

11-15 Filled with 0s since SORTJNF2 does not participate in

building this record.

Table 37. Record Format after REFORMAT Processing

For more examples, see “Joining Records from Multiple Files” on page 3-14.

Syncsort MFX Programmer’s Guide 2–233

SORT

SORT Control Statement

The SORT control statement defines the application as a sort or copy application.
Either a SORT control statement or a MERGE control statement is required for
every application.

Cultural Environment Support

Cultural environment support allows you to choose an alternative set of collating
rules based on a specified national language. The alternative collating applies to
SORT/MERGE and INCLUDE/OMIT processing.
For additional detail, see “LOCALE” on page 5-18.

SORT Control Statement Format

The format of the SORT control statement is illustrated below:

 FIELDS=(p 1 ,l 1 [ ,f 1 ],o 1 [,p 2 ,l 2 [ ,f 2 ],o 2 ]...)[,FORMAT=f] 

 
SORT  FIELDS=COPY 

0 
  ,CKPT
,CENTWIN=  s 
f  ,CHKPT
 

d 
 
  ( nn,mm )  
,DYNALLOC =  (d,n (,RETRY=  OFF  [ ,SC=s ]) 
   
 
 OFF 

,EQUALS n  n 
,FILSZ=   ,SIZE=  
,NOEQUALS  En   En 

[ ,SKIPREC=n ] [ ,STOPAFT=n ]

Figure 139. SORT Control Statement Format

FIELDS Parameter (Required)

The FIELDS parameter is required. It describes the control fields.
List the control fields in order of greatest to least priority, with the primary control
field listed first, followed by progressively less significant fields. You can specify up
to 128 control fields; however, if fields are complex, the limit for a particular
execution may be less than 128.

2–234 Syncsort MFX Programmer’s Guide

 SORT

Each field specified in the FIELDS parameter is identified by its position (p),
length (l), format (f) and order (o).
p The position value indicates the first byte of the field relative to the begin-
ning of the input record after INREC and/or E15 processing, if specified,
have completed.
Binary control fields can begin on any bit of a byte. When a binary field does
not begin on a byte boundary, you must specify the bit number (0-7). For
example, a position value of 21.3 refers to the 4th bit of the 21st byte of the
record.
l The length value indicates the length of the control field. The length value
must be an integer number of bytes except for the length of a binary control
field which can be specified in bits. For example, a length value of 0.5 refers
to a binary control field 5 bits long.
For signed fields, the length value must include the area occupied by the
sign.
f The format value indicates the data format. For a list of valid formats, refer
to the table in the next section, “Valid Formats for Sort Control Fields.” If all
the control fields have the same format, you can specify the format value
once by using the FORMAT=f subparameter. If you specify both the individ-
ual f values and the FORMAT subparameter, the individual f values will be
used for fields where they are specified.
o The order value indicates how the field is to be collated:
• A=Ascending order
• D=Descending order
• E=As modified by an E61 exit. Ascending order

Valid Formats for Sort Control Fields

The following chart lists the valid formats for sort control fields.

Syncsort MFX Programmer’s Guide 2–235

SORT

Field Length
Code Data Format
(bytes)

AC EBCDIC characters are translated to their ASCII equivalents 1 to 4091†

before sorting.

AQ Character. Records are sorted according to an alternate sequence 1 to 4091†

specified either in the ALTSEQ control statement or as an installa-
tion default.

ASL Leading separate sign. An ASCII + or - precedes numeric field. One 2 to 256
digit per byte.

AST Trailing separate sign. An ASCII + or - trails numeric field. One 2 to 256
digit per byte.

BI Binary. Unsigned. 1 bit to 4092*

CH Character. Unsigned. 1 to 4092*

CLO Leading overpunch sign. Hexadecimal F,C,E, or A in the first 4 bits 1 to 256
OL of your field indicates a positive number. Hexadecimal D or B in the
first 4 bits indicates a negative number. One digit per byte.
CMP=CLC is forced.

CSF Floating sign format. An optional leading sign may be specified 1 to 32

FS immediately to the left of the digits. If the sign is a -, the number is
treated as negative. For other characters, the number is treated as
positive. Characters to the left of the sign are ignored.

CSL Leading separate sign. An EBCDIC + or - precedes numeric field. 2 to 256

LS One digit per byte. CMP=CLC is forced.

CST Trailing separate sign. An EBCDIC + or - follows numeric field. 2 to 256

TS One digit per byte. CMP=CLC is forced.

FD Decimal floating point. Signed. An SNaN or QNaN value is invalid 4, 8, or 16

and will cause a WER497A error.

FI Fixed point. Signed. (Equivalent to Signed Binary.) 1 to 256

FL Floating point. Normalized. Signed. 2 to 256

PD Packed decimal. Signed. 1 to 256

PD0 Packed decimal. 2-8-byte packed decimal data with the first digit 2-8
and trailing sign ignored. The remaining bytes are treated as
packed decimal digits. Typically PD0 is used with century window
processing and Y2P format; Y2P processes the year, while PD0 pro-
cesses month and day.

Table 38. (Page 1 of 3) Format Code Chart

2–236 Syncsort MFX Programmer’s Guide

 SORT

Field Length
Code Data Format
(bytes)

SFF Signed free format. Decimal digits (0-9) are extracted from right to 1 to 44
left to form a number value. A character of – or ) found within the
field will cause the value to be treated as a negative number. All
other non-decimal digit values in the field are ignored.

UFF Unsigned free format. Decimal digits (0-9) are extracted from right 1 to 44
to left to form a number value. All non-decimal digit values in the
field are ignored.

Y2B Binary. 2-digit, 1-byte binary year data treated as a 4-digit year by 1
CENTWIN (century window) processing.

Y2C Character. 2-digit character year data treated as a 4-digit year by 2

CENTWIN (century window) processing. Processing is identical to
Y2Z fields.

Y2D Packed decimal. 2-digit, 1-byte packed decimal year data treated as 1
a 4-digit year by CENTWIN (century window) processing.

Y2P Packed decimal. 2-digit, 2-byte packed decimal year data. Of the 2
four packed digits contained in the 2 bytes, the first digit and trail-
ing sign are ignored; the two inner digits are treated as a 4-digit
year by CENTWIN processing.

Y2S Character or zoned decimal. 2-digit, 2-byte valid numeric data 2

treated as a 4-digit year by CENTWIN (century window) process-
ing, as for Y2C and Y2Z. However, certain data are not treated as
year data. Data with binary zeros (X'00') or a blank (X'40') in the
first byte will be collated before valid numeric year data for ascend-
ing order (after year data for descending order). Data with all
binary ones (X'FF') in the first byte will be collated after valid
numeric year data for ascending order (before year data for
descending order). Zones are ignored, as for Y2C and Y2Z, except
for data where the first byte begins with X'00', X'40' or X'FF'.

Y2T Full-date, character, binary, or packed decimal formats. Full-date 2-6

data formats can be used to sort or merge a variety of date fields.
Y2U They can process dates ending or starting with year digits (x...xyy
or yyx...x). They can also process non-date data commonly used
Y2V
with dates. For details, see Table 40 on page 2-244.
Y2W

Y2X

Y2Y

Y2Z Zoned decimal. 2-digit, 2-byte zoned decimal year data treated as a 2
4-digit year by CENTWIN (century window) processing. The zones
are ignored. Processing is identical to Y2C fields.

Table 38. (Page 2 of 3) Format Code Chart

Syncsort MFX Programmer’s Guide 2–237

SORT

Field Length
Code Data Format
(bytes)

ZD Zoned decimal. Trailing overpunch in the first 4 bits of the right- 1 to 256
CTO most byte gives the sign. Hexadecimal F,C,E, or A indicates a posi-
OT tive number. Hexadecimal D or B indicates a negative number. One
digit per byte. CTO forces CMP=CLC.

Note: * 4084 for variable-length records.

† 2043 for variable-length records.

Table 38. (Page 3 of 3) Format Code Chart

For information on the year data formats (Y2B, Y2C, Y2D, Y2P, Y2S, and Y2Z) plus
the related data format PD0 and the full-date formats, see “CENTWIN Parameter
(Optional)” on page 2-239, “Converting Year Data with Century Window Processing
on INREC, OUTREC, or OUTFIL OUTREC” on page 2-162, and “Specifying Field-
to-Field Standard Comparisons for Year Fields” on page 2-39.

Rules for Specifying Sort Control Fields

• For fixed-length records, all control fields and the sum of their lengths cannot
exceed 4092 bytes. When EQUALS is in effect, the number is reduced 4 bytes to
4088 bytes. EXTCOUNT also reduces the number by 4 bytes. Thus, if both
EQUALS and EXTCOUNT are in effect, the number is reduced to 4084 bytes.
• For variable-length records, all control fields must be located within the first
32750 bytes and the sum of their lengths cannot exceed 4084 bytes. When
EQUALS is in effect, all control fields must be located within the first 32746
bytes and the sum of their lengths cannot exceed 4080 bytes.
• Control fields can be in contiguous or non-contiguous locations in the record.
• Remember that for variable-length records, the first 4 bytes are reserved for the
Record Descriptor Word, so the first byte of the data portion of the record is byte
5.
• If the output file is a key-sequenced VSAM cluster, the VSAM key must be the
first control field specified.

Comparing PD and ZD Control Fields

The CMP PARM determines how PD and ZD control fields will be compared. When
CMP=CPD is in effect, the Compare Decimal (CP) instruction may be used under
certain circumstances for the compare. ZD fields are packed and then compared.
This method has performance advantages. However, invalid PD data may cause a
system 0C7 abend and program termination. Moreover, the integrity of ZD fields is
only guaranteed when they contain valid ZD data. The CMP=CPD method will not
be used for control fields that exceed 16 bytes or for variable-length merges when
an even value (0, 2, 4, or 6) is specified for the VLTEST PARM.

2–238 Syncsort MFX Programmer’s Guide

 SORT

When CMP=CLC is in effect, no data validation is performed and the integrity of

the output is maintained, even if the sign for a PD or ZD field is invalid. This
method will be used if any control field exceeds 16 bytes or for variable-length
merges when an even value is specified for the VLTEST PARM.

CENTWIN Parameter (Optional)

The CENTWIN run-time or installation option acts on 2-digit year data.
CENTWIN generates a century window (for example, 1950 through 2049) that
determines the century to which a 2-digit year belongs. At run-time, CENTWIN
can be specified as either a PARM option or a SORT/MERGE control statement
parameter. CENTWIN ensures that year data spanning centuries will be
sequenced correctly. Without CENTWIN processing, an ascending sort would
sequence the year 01 before the year 98. With CENTWIN processing, the 01 field
could be recognized as a twenty-first century date (2001) and would thus be
sequenced after 98 (1998).
For more information on specifying the CENTWIN option, see “CENTWIN” on
page 5-6.
CENTWIN SORT/MERGE processing only applies to data defined as year data
formats: Y2B, Y2C, Y2D, Y2P, Y2S, Y2Z, and the full-date formats (Y2T, Y2U, Y2V,
Y2W, Y2X, and Y2Y). These data formats enable MFX to process 2-digit year fields
as 4-digit years. A related data format, PD0, can be used to process the month and
day portions of packed decimal date fields. To correctly specify date fields for
CENTWIN SORT processing, you should be familiar with the CENTWIN-related
data formats.
The following describes each of the year data formats and provides SORT control
statement examples:
The Y2B Format
This format is used to sequence 2-digit, 1-byte binary year data with CENTWIN
processing. The binary values are converted to decimal, and the two low order
digits are used as year data. Thus, while binary and decimal values range from 00
to 255, year values range from 00 to 99. The relationship between binary, decimal
and year values is shown in the following table:

Binary Value Decimal Value Year Value

X'00' to X'63' 00 to 99 00-99

X'64' to X'C7' 100 to 199 00-99

X'C8' to X'FF' 200 to 255 00-55

Table 39. Possible Values Representing Year Data with Y2B

Syncsort MFX Programmer’s Guide 2–239

SORT

The Y2C and Y2Z Formats

These formats represent 2-digit, 2-byte year data in either character (Y2C) or
zoned decimal (Y2Z) format. Either Y2C and Y2Z formats can be used with data of
the form
X'xyxy'
where y is a hexadecimal year digit 0-9 and x is hexadecimal 0 through F. Y2C and
Y2Z ignore the x digits, leaving yy, the 2-digit unsigned year representation.
Suppose you have a character or zoned decimal date field mmddyy that begins at
byte 20. You can use either Y2C or Y2Z to process the yy field. As the following
example indicates, you could specify three sort keys to correctly sort this date:

SORT FIELDS=(24,2,Y2C,A, * Sorts yy field as 4-digit year

20,2,CH,A, * Sorts mm field
22,2,CH,A) * Sorts dd field

Figure 140. Sample SORT Statement

The yy field (24,2) will be processed according to the century window setting. For
example, if CENTWIN=1945, the field yy=45 will be sequenced as if it were 1945,
and yy=44 would be sequenced as if it were 2044. Thus, for an ascending sort, 44
would follow 45.
The Y2D Format
This format is used to sequence 2-digit, 1-byte packed decimal year data with
CENTWIN processing. Use Y2D to extract the year data yy from packed decimal
date fields. For example, consider a 3-byte packed decimal data field defined as
X'yyddds'
This field has the year yy in the first byte and the day ddd in bytes 2 and 3. The
packed decimal sign s would be in the last digit (half byte) of the third byte. To sort
this date field, which begins at byte 20, with 4-digit year processing, use the
following SORT control statement:

SORT FIELDS=(20,1,Y2D,A, * Sorts 2-digit year (yy) as 4-digit year

21,2,PD,A) * Sorts ddds as 3 digits (ddd)

Figure 141. Sample SORT Statement

The Y2P Format

This format is used to sequence 2-digit, 2-byte packed decimal year data with
CENTWIN processing. Use Y2P to extract the year data yy from packed decimal
date fields spanning 2 bytes. For example, a packed decimal date of the form
yymmdd would be stored as 4 bytes:

2–240 Syncsort MFX Programmer’s Guide

 SORT

yymmdd = X'0yymmddC'
where the trailing C (sometimes F) is a positive sign and the leading 0 pads the
field on the left to make an even number of digits.
Notice that the components of the date span bytes:
0y ym md dC
Y2P handles this condition by ignoring the first and last half bytes of the 2-byte
field specification. Thus, Y2P processes 0yym as yy, ignoring the leading digit (0)
and the trailing digit m that is part of the month.
The following example uses Y2P to sort the year portion of the date field, which
begins at byte 20:

SORT FIELDS=(20,2,Y2P,A) * Sorts yy field as 4-digit year

Figure 142. Sample SORT Statement Using Y2P

The field specification 20,2,Y2P treats X'0yym' as X'yy', and CENTWIN processing
sorts yy as a 4-digit year yyyy.
The PD0 format, described below, can assist Y2P by processing month and day data
that overlap year data in the original field.
The Y2S Format
This format is used to sequence 2-digit, 2-byte character or zoned decimal data. The
Y2S format is identical to Y2C and Y2Z for valid numeric data, but Y2S treats data
that begin with X'00', X'40', or X'FF' as non-year data. Thus, the Y2S format can
distinguish records that have non-year data in the first byte of the year field,
allowing such records to be sorted differently from other records.
Y2S treats non-year data as follows:
• Data with binary zeros (X'00') or a blank (X'40') in the first byte will not have
century window processing applied to it. Instead, such data will be collated in
sequence, before valid numeric year data for ascending order or after the year
data for descending order.
• Data with all binary ones (X'FF') in the first byte will also not have century
window processing applied to it. Instead, such data will be collated after valid
year numeric data for ascending order or before the year data for descending
order.
• Zones are ignored, as for Y2C and Y2Z, except for data where the first byte
begins with X'00', X'40', or X'FF'.
As an example, suppose you want to preserve the input order of header and trailer
records at the start or end of the file, and your header/trailer records are identified
by binary zeros (X'00'), a blank (X'40'), or binary ones (X'FF') in the first byte of the
date field.

Syncsort MFX Programmer’s Guide 2–241

SORT

The Y2S format allows CENTWIN to identify the header/trailer records and treat
them differently from other records. Presuming the year data begin in column 20,
you would use the following sort key specification:

SORT FIELDS=(20,2,Y2S,A) * Sorts yy field as 4-digit year

Figure 143. Sample SORT Statement

The yy field (20,2) will be processed according to the century window setting. For
CENTWIN=1945, data with header and trailer records would be sorted as follows:

SORTIN Input Record Order after Sorting

X'F9F6' X'0000'

X'4001' X'4000'

X'F4F4' X'4001'

X'4000' X'F5F1'

X'0000' X'F9F6'

X'F5F1' X'F4F4'

X'FF03' X'FF03'

Note that if the above data were sorted as Y2C or Y2Z format, the output order
would be different because the records starting with X'00', X'40', and X'FF' would
be interpreted as numeric years. For example, suppose the fields in the above list
were defined as Y2Z and sorted with EQUALS:

SORT FIELDS=(20,2,Y2Z,A),EQUALS

Figure 144. Sample SORT Statement

The data would be processed as follows:

SORTIN Input Record Order after Sorting

X'F9F6' X'F5F1'

X'4001' X'F9F6'

X'F4F4' X'FF03' (invalid numeric data)

X'4000' X'4000' (invalid numeric data)

X'0000' X'0000' (invalid numeric data)

2–242 Syncsort MFX Programmer’s Guide

 SORT

X'F5F1' X'4001' (invalid numeric data)

X'FF03' X'F4F4'

The header and trailer records are sequenced as year data according to the
CENTWIN setting (CENTWIN=1945), and they lose their position at the start and
end of the file.
The PD0 Format
This format is used to sequence 2-8 byte packed decimal data. PD0 ignores the first
digit and trailing sign during processing. PD0 is normally used in conjunction with
the Y2P data format. The Y2P format is used to process the 2-digit year portion of a
packed decimal date field, while the PD0 format is used to process the month and
day portion of the field.
Although PD0 is typically used with Y2P, the PD0 format itself is not affected by
CENTWIN processing.
Consider the packed decimal date field used in the example above:
yymmdd = X'0yymmddC'
where the trailing C (sometimes F) is a positive sign and the leading 0 pads the
field on the left to make an even number of digits.
Notice that the components of the date span bytes:
0y ym md dC
The date can be processed as follows:
• Y2P processes the year component X'0yym' as X'yy'.
• PD0 processes the month and day components X'ymmddC' as X'mmdd'.
The following SORT control statement can be used to sort the entire date with
CENTWIN processing:

SORT FIELDS=(20,2,Y2P,A, * Treats X'0yym' as X'yy'; sorts yy as yyyy

21,3,PD0,A) * Treats X'ymmddC' as X'mmdd'

Figure 145. Sample SORT Statement

Full-Date Formats
Full-date formats can be used to sort or merge various date fields, processing dates
ending or starting with year digits. They also process non-date data that are used
with dates. For a full description of full-date formats, see the following section.

Syncsort MFX Programmer’s Guide 2–243

SORT

Using Full-Date Formats with CENTWIN

MFX’s full-date data formats enable you to sort or merge a variety of date fields.
The full-date formats are Y2T, Y2U, Y2V, Y2W, Y2X, and Y2Y. These date formats
can process dates ending or starting with year digits:
• x...xyy (for example: qyy, mmyy, dddyy, or mmddyy)
• yyx...x (for example: yyq, yymm, yyddd, or yymmdd)
The full-date formats also process non-date data commonly used with the dates.
MFX interprets two-digit years (yy) according to the century window specified by
the CENTWIN option. CENTWIN processing does not apply to non-date data.
In most cases, for CH, ZD, and PD date fields the full-date data formats are easier
to use than the 2-digit date formats. The 2-digit formats can be more difficult
because you must divide the date into its components. This requires care,
particularly for PD dates, where date components (q, dd, mm, or yy) may span
bytes or occupy only part of a byte. The full-date formats, on the other hand,
process such dates automatically.
The table below describes the full-date formats. For date forms not in the table, use
the 2-digit year formats or the non-year formats.
Note the following symbols used in the table:
y year digit (0-9)
x non-year digit (0-9)
s sign (hexadecimal A-F)
0 unused digit

Full-Date Data Example Date

Date Form Length (bytes)
Format Format Form

Y2T CH, BI yyx yyq 3

yyxx yymm 4

yyxxx yyddd 5

yyxxxx yymmdd 6

Y2U PD yyx yyq 2

(X'yyxs')

yyxxx yyddd 3

(X'yyxxxs')

Table 40. Full-Date Formats

2–244 Syncsort MFX Programmer’s Guide

 SORT

Full-Date Data Example Date

Date Form Length (bytes)
Format Format Form

Y2V PD yyxx yymm 3

(X'0yyxxs')

yyxxxx yymmdd 4
(X'0yyxxxxs')

Y2W CH, BI xyy qyy 3

xxyy mmyy 4

xxxyy dddyy 5

xxxxyy mmddyy 6

Y2X PD xyy qyy 2

(X'xyys')

xxxyy dddyy 3

(X'xxxyys')

Y2Y PD xxyy mmyy 3

(X'0xxyys')

xxxxyy mmddyy 4
(X'0xxxxyys')

Table 40. Full-Date Formats

The table indicates the full-date formats that can be used with character (CH),
binary (BI), or packed decimal (PD) data. Note the recognized non-date values:
Character or binary (Y2T and Y2W full-date formats)
C'0...0' (CH zeros)
C'9...9' (CH nines)
Z'0...0' (ZD zeros)
Z'9...9' (ZD nines)
X'00...00' (BI zeros)
X'40...40' (blanks)
X'FF...FF' (BI ones)
Packed (Y2U, Y2V, Y2X, and Y2Y full-date formats)
P'0...0' (PD zeros)
P'9...9' (PD nines)
The following two examples illustrate how you might use Table 40 (“Full-Date
Formats”) on page 2-244:
• Suppose you have a packed decimal (PD) date field of the form mmyy. To sort
this field correctly, you would use the Y2Y 3-byte format from the table. Thus, if

Syncsort MFX Programmer’s Guide 2–245

SORT

the field starts in position 30, you would specify the following SORT control
statement to sort in descending order:
SORT FIELDS=(30,3,Y2Y,D)
Any PD fields of all PD zeros or all PD nines will be processed automatically as
non-date data.
• Suppose you have a character (CH) date field of the form yymmdd. To sort this
field correctly, you would use the Y2T 6-byte format from the table. Thus, if the
field starts in byte 40, you would specify the following SORT control statement
to sort in ascending order:
SORT FIELDS=(40,6,Y2T,A)
Any CH zeros, CH nines, BI zeros, blanks, and BI ones will be processed
automatically as non-date data.
Collating Sequence with Full-Date Formats
For full-date formats, the yy component is always sorted first (treated as primary
key). This is so even when the yy is physically at the rightmost end of the field, as
for Y2W, Y2X, and Y2Y. For example, a 6-byte Y2W field has the form xxxxyy. This
is collated with the yy as the primary key and xxxx as the secondary key. Because
MFX automatically collates the year character first, you don’t have to deal with yy
manually, for example by using PD0 and Y2D.
It is important to understand that the xxxx component of a full-date format must
be designed to collate as a unit. Suppose you have the 6-byte Y2T field yyxxxx. If
you collate this field in ascending order, then yy collates first (the primary key)
with xxxx collating second (secondary key). Consider two possibilities:
• If yyxxxx is actually yymmdd, you will be sorting first by year, then month, then
day.
• If yyxxxx is actually yyddmm, you will sorting by year, then day, then month. In
most cases, sorting in this way would not be what you intended.
To correctly collate a date, the date components must be in an order suitable for
collating. For example, mmddyy and yymmdd will collate correctly, but ddmmyy or
yyddmm will not. For date forms that will not collate correctly, you must use one of
the 2-digit year formats (Y2B, Y2C, Y2D, Y2P, Y2S, and Y2Z).

2–246 Syncsort MFX Programmer’s Guide

 SORT

The following table shows the order for ascending collation when using full-date
formats with the CENTWIN option:

Full-Date Format Date Format Ascending Sort Sequence

Y2T CH, BI BI zeros

Y2W Blanks

CH/ZD zeros

Lower century dates (e.g. 1980)

Higher century dates (e.g. 2010)

CH/ZD nines

BI ones

Y2U PD PD zeros

Y2V Lower century dates (e.g. 1980)

Y2X Higher century dates (e.g. 2010)

Y2Y PD nines

Table 41. Full-Date Formats, Ascending Collation

For a descending sort, the collation order is reversed.

Other date formats (non-full-date), with the exception of Y2S, do not process
non-date data; their sort sequence for ascending sorts is simply lower century
dates than higher century dates.

Examples Using Full-Date Formats

Example 1 (Y2W)
The following SORT control statement sorts a C'mmddyy' date field in ascending
order, with the previously set fixed century window 1984-2083:
SORT FIELDS=(10,6,Y2W,A) * Sort C'mmddyy' in ascending order
* with Y2W
* and previously set century window 1984-2083
The Full-Date Formats table above indicates that the 6-byte Y2W form is
appropriate for a CH input field of the form xxxxyy. As shown in the following table,

Syncsort MFX Programmer’s Guide 2–247

SORT

the output will be collated as C'yyyymmdd', with the non-date data (zeros)
appearing correctly at the beginning of the sorted output.

SORTIN Input Record Order Actual Date

mmddyy
after Sorting after Sorting

mmddyy yyyy/mm/dd

021783 000000 non-date data

092206 070484 1984/07/04

081395 081395 1995/08/13

110210 092206 2006/09/22

000000 110210 2010/11/02

070484 043060 2060/04/30

043060 021783 2083/02/17

Example 2 (Y2T)
The following SORT control statement sorts a Z'yyddd' date field in descending
order, with the previously set fixed century window 1921-2020:
SORT FIELDS=(20,5,Y2T,D) * Sort Z'yyddd' in descending order
* with Y2T
* and previously set century window 1921-2020
The Full-Date Formats table above indicates that the 5-byte Y2T form is
appropriate for a ZD input field of the form yyddd. As shown in the following table,
the output will be collated as Z'yyyyddd', with the non-date data (nines and zeros)
appearing correctly at the beginning and end of the sorted output.

SORTIN Record Order Actual Date

Input after Sorting after Sorting

yyddd yyddd yyyy/ddd

00000 99999 non-date data

50237 20153 2020/153

99999 20047 2020/047

20047 01223 2001/223

94001 94001 1994/001

01223 50237 1950/237

2–248 Syncsort MFX Programmer’s Guide

 SORT

20153 21148 1921/148

21148 00000 non-date data

Example 3 (Y2Y)
The following SORT control statement sorts a P'mmddyy' (X'0mmddyys') date field
in ascending order, with the previously set fixed century window 1921-2020:
SORT FIELDS=(26,4,Y2Y,A) * Sort P'mmddyy' in ascending order
* with Y2Y
* and previously set century window 1921-2020
The Full-Date Formats table above indicates that the 4-byte Y2Y form is
appropriate for a PD input field of the form xxxxyy. As shown in the following table,
the output will be collated as P'yyyymmdd', with the non-date data (zeros and
nines) appearing correctly at the beginning of the sorted output. Note that the first
two columns are in hexadecimal.

SORTIN Record Order Actual Date

Input after Sorting after Sorting

mmddyy mmddyy yyyy/mm/dd

0999999C 0000000C non-date data

0102250C 0080321C 1921/08/03

0032120C 0102250C 1950/10/22

0010194C 0010194C 1994/01/01

0000000C 0111501C 2001/11/15

0111501C 0032120C 2020/03/21

0080321C 0999999C non-date data

FIELDS=COPY (Required for a Copy)

Use FIELDS=COPY to copy one or more input files. Multiple files can be copied if
they are concatenated to the SORTIN DD statement. Other control statements
such as INREC, INCLUDE/OMIT, OUTREC, and OUTFIL may be specified in
conjunction with a copy application, allowing you to edit and reformat the file(s)
without sorting them.
The SUM or DUPKEYS control statement and an E32 exit cannot be specified with
FIELDS=COPY. All Phase 3 exits can be used.

Syncsort MFX Programmer’s Guide 2–249

SORT

CKPT/CHKPT Parameter (Optional)

The CKPT/CHKPT parameter instructs MFX to take a checkpoint at every end-of-
volume of a SORTOUT data set when OUTFIL is not used and also at the
beginning of Phase 3 before the SORTOUT data set is opened. Either spelling of
this parameter is accepted.
This parameter requires a SORTCKPT DD statement. It cannot be specified in
conjunction with a user-issued STIMER macro or an incore sort. Checkpoints
cannot be taken within a user exit routine.
Refer to “Chapter 14 Performance Considerations” for an explanation of the
Checkpoint/Restart feature.

DYNALLOC Parameter (Optional)

The format of the DYNALLOC parameter is illustrated below.

 d 
 
   
,DYNALLOC =  ( d,n ,RETRY =  ( nn,mm )  [,SC = s ) 
  OFF  
 
 OFF 

Figure 146. DYNALLOC Parameter Format

DYNALLOC requests the dynamic allocation of SORTWK data sets on device type
d. Specify the device type either as a decimal number (e.g., 3390) or by the system
generic name (e.g., SYSDA). Any disk device accepted for a SORTWK DD
statement can be specified. Note that if VIO is specified it will be ignored, and the
installation default for the DYNALLOC device type will be used in its place.
Note that the DYNALLOC parameter may be used alone, without any
subparameters. In this case, the DYNALLOC installation default settings are used.
For MAXSORT applications, n is the number of SORTWK data sets that will be
allocated. As many as 32 SORTWK data sets can be specified. The default for n is 3.
For non-MAXSORT applications, n can be 1 through 255. This value specifies the
number of SORTWK data sets that can potentially be allocated. For values of n
that are 31 or less, MFX can automatically raise the number to 32 if the application
requires it. When n is 33 through 255, this value specifies the maximum number of
SORTWK data sets that can be allocated.
DYNALLOC=OFF can be specified to override a DYNALLOC=ON installation
default.
Normally for both MAXSORT and non-MAXSORT applications, any SORTWK data
sets provided in the JCL will contribute towards the value of n. For instance, if n
was set to 40 in a non-MAXSORT application and 30 SORTWKs were provided in

2–250 Syncsort MFX Programmer’s Guide

 SORT

the JCL, DYNALLOC could obtain 10 additional SORTWKs if needed. Note that
there is an installation option to disable DYNALLOC if SORTWKxx DD
statements are present.
MFX uses the value specified in the RETRY parameter to request automatic
DYNALLOC retry. This facility attempts to avoid a sortwork capacity exceeded
condition when disk space is not immediately available to satisfy a DYNALLOC
request. MFX will automatically retry a specified number of times and wait a
prescribed interval between DYNALLOC requests.
The nn in the first position designates the number of times MFX will retry a failed
DYNALLOC request. The minimum allowed is 0 and the maximum is 16. The mm
in the second position designates the number of minutes MFX waits between each
DYNALLOC request. The minimum allowed is 0 and the maximum is 15. A value
of 0 can be used to request an immediate retry. RETRY=OFF or an nn of 0 can be
specified to override a RETRY=ON installation default.
In an environment where DFSMS manages temporary work data sets, the SC
subparameter specifies a storage class s for MFX to use when dynamically
allocating SORTWORK data sets. The storage administrator at your installation
defines the names of the storage classes you can specify. Note that an installation
written automatic class selection (ACS) routine can override the storage class you
specify. If SMS is not installed or active to manage temporary work data sets, the d
device specification will be used in the SORTWORK dynalloc request.

EQUALS/NOEQUALS Parameter (Optional)

The EQUALS parameter insures that the original order of equal-keyed records is
preserved. These records will be in the same order in the output file as they were in
the input file. NOEQUALS, the default, specifies that equal-keyed records may not
be written in their original input order.
If EQUALS is in effect in an application with SORTMInn data sets, the order of
equal-keyed records within each SORTMInn file will be preserved. In addition,
equal-keyed records from the lowest-numbered SORTMInn file will be written
before those from the second SORTMInn file, and so on.
When the EQUALS parameter is used with the SUM or DUPKEYS control
statement, the first of the equal-keyed records is retained with the sum or
DUPKEYS function value; all other records are deleted after the specified field(s)
have been calculated.
EQUALS/NOEQUALS can also be specified as a PARM option on the EXEC
statement. If this option is specified both on the SORT control statement and as a
PARM option, the SORT specification takes precedence.
Performance is usually improved when NOEQUALS is in effect.

Syncsort MFX Programmer’s Guide 2–251

SORT

FILSZ Parameter (Optional)

The FILSZ parameter specifies the actual (FILSZ=n) or estimated (FILSZ=En)
decimal number of records to be sorted. This number should reflect any changes
produced by INCLUDE/OMIT, E14 and/or E15, SKIPREC and STOPAFT
processing.
If FILSZ=n is specified, MFX will terminate unless exactly n records are processed.
FILSZ can also be specified as a PARM option on the EXEC statement. If this
option is specified both on the SORT control statement and as a PARM option, the
PARM specification takes precedence.

SIZE Parameter (Optional)

The SIZE parameter specifies the actual (SIZE=n) or estimated (SIZE=En) decimal
number of records read from the input file. Unlike the FILSZ parameter, this
number should not reflect any changes produced by INCLUDE/OMIT or exit
processing, but should reflect SKIPREC and STOPAFT processing.
If the FILSZ parameter is not specified and SIZE=n is specified, MFX will
terminate unless exactly n records are processed. If the FILSZ parameter is
specified, the SIZE value is considered an estimate whether or not it is preceded by
an E.

SKIPREC Parameter (Optional)

The SKIPREC=n parameter instructs MFX to skip a decimal number of records
before the input file is sorted or copied. The n records skipped are deleted from the
input file before E15 and INCLUDE/OMIT processing, if specified, take place.
If SKIPREC is specified as a PARM option as well as on the SORT control
statement, the PARM specification takes precedence.

STOPAFT Parameter (Optional)

The STOPAFT=n parameter specifies the number of records to be sorted or copied.
These will be the first n records after E15, INCLUDE/OMIT and SKIPREC
processing, if specified, have completed. If STOPAFT is specified as a PARM option
as well as on the SORT control statement, the PARM specification takes
precedence.

Sample SORT Control Statements

SORT FIELDS=(2.3,2,BI,D,8,2.4,BI,A,25,10,CH,A,15,10,LS,D)

Figure 147. Sample SORT Control Statement

This sample SORT control statement indicates four control fields:

2–252 Syncsort MFX Programmer’s Guide

 SORT

• The first, or primary, field begins in bit 4 of byte 2, is 2 bytes long, is in binary
format and is to be sorted in descending order.
• The second control field begins in byte 8, is 2 bytes 4 bits long, is a binary
format and is to be sorted in ascending order.
• The third control field begins on byte 25, is 10 bytes long, is in character format
and is to be sorted in ascending order.
• The fourth control field begins on byte 15, is 10 bytes long, is an EBCDIC
numeric field with a leading separate sign and is to be sorted in descending
order.

SORT FIELDS=(20,5,A,5,10,D,30,5,A),FORMAT=CH,CKPT

Figure 148. Sample SORT Control Statement

This sample SORT control statement specifies the following:

• There are three control fields. Because all three fields have the same data
format (in this case, character), the FORMAT=CH subparameter is specified so
that the CH value does not have to be specified for each of the fields.
• The first control field begins on byte 20, is 5 bytes long and is to be sorted in
ascending order.
• The second control field begins on byte 5, is 10 bytes long and is to be sorted in
descending order.
• The third control field begins on byte 30, is 5 bytes long and is to be sorted in
ascending order.
• MFX will take a checkpoint.

Syncsort MFX Programmer’s Guide 2–253

SUM

SUM Control Statement

The SUM control statement allows you to sum numeric fields in records with equal
sort/merge keys, place the sum in one record which is retained, and delete the other
equally-keyed records. Provided arithmetic overflow does not occur during the
summing process, the SUM control statement produces only one record per
sort/merge key. The records deleted by SUM can optionally be written to a separate
output file.
SUM FIELDS=NONE can be used to delete all but one of the records with equal
keys without doing any summing.
SUM can also be specified on the DUPKEYS control statement to perform the same
function. The DUPKEYS statement provides additional functions for equally-keyed
records such as providing AVG, MAX and MIN values.
The SUM control statement should not be used and will be ignored when
FIELDS=COPY is specified on the SORT or MERGE control statement.

SUM Control Statement Format

The format of the SUM control statement is illustrated below.

 FIELDS=(p 1 ,l 1 [ ,f 1 ][,p 2 ,l 2 [ ,f 2 ]]...)[,FORMAT=f] 

SUM  FIELDS=NONE [,XSUM]
 

Figure 149. SUM Control Statement Format

FIELDS Parameter (Required)

The FIELDS parameter defines the numeric fields to be summed when the control
fields of two or more records are equal. Specify FIELDS=NONE to reduce the
sorted data to one record per sort key without summing any numeric fields.
Each field specified in the FIELDS parameter is identified by its position p, length
l and format f.
p The position value indicates the first byte of the field relative to the
beginning of the input record after INREC and/or E15 processing, if
specified, have completed. The field must begin on a byte boundary.
l The length value indicates the length of the field. The length must be
an integer number of bytes. Refer to Table 42 for the permissible
lengths.
f The format value indicates the data format. Table 42 lists the valid
data formats for SUM fields. If all the summed fields have the same
format, you can specify the format value once by using the FOR-

2–254 Syncsort MFX Programmer’s Guide

 SUM

MAT=f subparameter. If both the individual f values and the FOR-

MAT subparameter are specified, the individual f values will be used
for fields where they are specified.

FORMAT
PERMISSIBLE LENGTH
CODE

BI 2, 4, or 8 bytes

FD* 4, 8, or 16 bytes

FI 2, 4, or 8 bytes

FL 4, 8, or 16 bytes

PD 1 to 16 bytes

ZD 1 to 31 bytes

Note: *A non-finite number in the data will cause a WER497A error.

Table 42. Permissible Formats and Lengths for SUM Fields

XSUM Parameter (Optional)

Specify the XSUM parameter if you want records deleted by SUM processing to be
written to a data set defined by the SORTXSUM DD statement. These records will
be written to SORTXSUM at the time of SUM processing. The records will not
undergo OUTREC, E35, and OUTFIL processing because such processing occurs
after SUM processing.
The DCB BLKSIZE of the SORTIN data set will not be used to determine the
BLKSIZE of the SORTXSUM data set. System determined blocksize will be used
when enabled and appropriate. Unblocked output will be generated if system
determined blocksize has been disabled and an explicitly specified blocksize has not
been provided in the JCL.
The XSUM file will be sequenced in the same order as the SORTOUT file.
Note that XSUM may increase system requirements:
• Adding XSUM to an existing sort application may result in an increase in the
amount of SORTWORK space required. This occurs because XSUM delays all
summing until Phase 3.
• Adding XSUM to an existing MAXSORT application could cause the generation
of additional intermediate output files (SORTOU00 or SORTOUnn). This
occurs because XSUM delays SUM processing until the final MAXSORT merge
pass.
• XSUM may require additional main memory. Specify a region size of 512K or
more.

Syncsort MFX Programmer’s Guide 2–255

SUM

General Considerations for SUM

• If NOEQUALS is in effect, the record which is retained is determined
arbitrarily. If EQUALS is in effect, the record which is retained is the first
record read in a SORT application; in a MERGE, the retained record will be
from the lowest-numbered input file. The EQUALS parameter can be specified
on the SORT or MERGE control statement or as a PARM option.
• A sort or merge control field cannot be summed. A portion of a control field
cannot be included in a sum field.
• Sum fields may not overlap each other.
• Non-sum fields remain unchanged and are retained from the record which
contains the sum.
• If arithmetic overflow or underflow occurs during the summing of two records,
those records are not summed and neither record is deleted. Further processing
is determined by the option selected at installation through the SUMOVFL
parameter or the run-time parameter OVFLO. If the RC16 option of this
parameter has been selected, processing will terminate with a WER049A
critical error. For the RC0 (the delivered default) or the RC4 option, sum
processing will continue and a WER049I message will be issued (only for the
first occurrence). If a subsequent pair of records with equal control fields can be
summed without causing overflow or underflow, they will be summed. To avoid
arithmetic overflow, use the INREC control statement to insert zeros of the
proper format immediately before the sum field. For example, for a PD field, use
nZ to insert binary zeros.
• Remember that the first 4 bytes of variable-length records are reserved for the
Record Descriptor Word, so the first byte of the data portion of the record is byte
5.
• SUM is incompatible with an incore sort. If you specify the SUM control
statement, allocate SORTWKxx data sets in the JCL or use the DYNALLOC
feature for dynamic SORTWK allocation. If no JCL SORTWKs are provided and
DYNALLOC is disabled by default, SUM will cause DYNALLOC to be enabled.
• When FL fields are summed, user-issued SPIE macros are not permitted and
exit routines must not produce exponent overflow or underflow. Because of the
numeric rounding performed by the hardware, the exact sum depends on the
order in which fields are summed. Thus, the sum may vary slightly for different
executions.
• By default, the sign byte of a positive summed ZD field will be converted to
printable format. If you want to disable this action, use the NZDPRINT PARM
option. Refer to “ZDPRINT” on page 5-33.

2–256 Syncsort MFX Programmer’s Guide

 SUM

Sample SUM Control Statements

The following SUM control statement eliminates equal-keyed records without
summing numeric fields. The XSUM option causes the eliminated records to be
written to a data set defined on the SORTXSUM DD statement.

SUM FIELDS=NONE,XSUM

Figure 150. Sample SUM Control Statement

Records with equal control fields will be eliminated from SORTOUT or SORTOFnn
data sets so that only one record is retained.
The following SUM control statement sums two numeric fields on records with
equal control fields.

SUM FIELDS=(20,4,32,4),FORMAT=PD

Figure 151. Sample SUM Control Statement

When the control fields are equal, this SUM control statement sums the numeric
data in the fields beginning in bytes 20 and 32. Because both fields are in packed
decimal format, the FORMAT=PD subparameter is used so that the PD value does
not have to be specified for each field.
Comprehensive examples illustrating the SUM control statement are provided in
“Chapter 3, How to Use the MFX Data Utility Features”.

Syncsort MFX Programmer’s Guide 2–257

SUM

2–258 Syncsort MFX Programmer’s Guide

Chapter 3 How to Use the MFX Data
Utility Features

This chapter assumes that you already know how to sort records and are ready to
use the MFX Data Utility features for any or all of the following:
• Selecting only those input records and data fields that are needed for an
application.
• Eliminating duplicate records.
• Consolidating records into a single record that contains the sum of any numeric
data fields.
• Joining records.
• Making output data printable and easy to read.
• Writing a multi-sectioned report complete with headers and trailers.
• Generating several output files and reports with a single pass of the sort.
• E-mailing a report in PDF format.
The following examples show how you can accomplish these tasks with MFX. Each
example is self-contained and provides coding instructions for both the required
JCL and the necessary control statements. Use them as starting points for your
own applications. For details of control statement syntax see “Chapter 2, MFX
Control Statements”.

Sample Data Utility Applications

The following chart lists applications that demonstrate the MFX features.

Syncsort MFX Programmer’s Guide 3–1

 Feature Application Page
Selecting Input Records Including Relevant Records 3.3
Omitting Irrelevant Records 3.5
Selecting Relevant Fields from the Selecting a Number of Fields from Longer Records 3.7
Input Records Eliminating Irrelevant Data Field(s) 3.8
Selecting Fields from Variable-Length Records 3.9
Combining Records within a File Combining Records and Summing Numeric Data Fields 3.12
Eliminating Duplicate Records 3.13
Joining Records from Multiple Files Joining Records 3.14
Retaining Unpaired Records from One of the Join Files 3.19
Retaining Unpaired Records from Both Join Files 3.21
Using Join to Copy a Large Number of Master File 3.23
Records
Making Output Records Printable Reordering the Positions of Record Fields 3.26
and Easy to Read Inserting Blanks and Repositioning Record Fields 3.27
Inserting Binary Zeros 3.29
Converting Unprintable Data to Readable Form 3.31
Converting Unprintable Data to Hexadecimal Format 3.33
Converting and Editing Unprintable Data 3.34
Putting a Data Field in Standard Format 3.36
Converting from Variable to Fixed-Length Format 3.38
Printing Input Records on Multiple Output Lines 3.39
Dividing a Report into Sections Dividing Output into Sections 3.41
Writing Headers and Trailers for a Writing a Title Page for a Report 3.43
Report Writing a Page Header 3.44
Writing a Section Header 3.46
Using a Header to Eliminate Duplication Information 3.47
within a Section
Writing a Report Trailer or Summary 3.49
Writing a Page Trailer 3.50
Totaling and Subtotaling Data Totaling Data at the End of a Report 3.51
Subtotaling Data at the End of a Page 3.53
Totaling Data at the End of a Section 3.54
Obtaining Maximum, Minimum and Printing Maximum, Minimum and Average Data in Sec- 3.57
Average Data tion Trailers
Counting Data Records Obtaining a Count of Data Records 3.59
Obtaining a Cumulative (Running) Count of Data 3.60
Records
Creating Multiple Output Files Generating Several Output Files with Different Informa- 3.63
tion
Writing Identical Output Files to Different Devices 3.65

E-mailing a Report in PDF Format Writing a Report, Creating PDF Output, and E-mailing 3.66
the Report to Recipients.

Table 43. MFX Feature Applications

3–2 Syncsort MFX Programmer’s Guide

Selecting Input Records
When only certain records from an input file are needed for an application, MFX
allows you to set up one or more logical conditions for including only those records.
Alternately, you may specify conditions for omitting records from an application.
Each condition is based on a comparison between two record fields or between a
record field and a constant. You may specify the constant as a positive or negative
decimal, a hexadecimal or binary constant, or a character literal. Multiple
conditions may be specified, provided you connect them with ANDs and ORs.
To specify the conditions for selecting records, use the INCLUDE/OMIT control
statement. For complete syntax, and examples of bit level criteria in record
selection, see “INCLUDE/OMIT Control Statement” on page 2-29
When processing variable-length records, by default all fields specified must be
contained within the record. If an application is expected to reference fields not
completely contained within the record, see “VLTESTI” on page 5-32. VLTESTI
provides for processing of records that do not contain all fields.

Including Relevant Records

Example: A school board requires a list of all students performing below their grade
level on standardized exams. (The record layout is given in Figure 152 and a
sample record is given in Figure 153.)

Figure 152. Input Record Layout

Syncsort MFX Programmer’s Guide 3–3

Figure 153. Sample Student Record

To generate the list, the following is coded:

//SUBLEV JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX
//* Messages to I/O Device
//SORTIN DD DSN=WWBRSM.STUDENTS,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD UNIT=SYSDA, Defines Intermediate
// SPACE=(CYL,15) Storage
//SYSIN DD *
INCLUDE COND=(29,2,LT,25,2,OR,27,2,LT,25,2),
FORMAT=PD Selects Records
SORT FIELDS=(1,14,CH,A) Sorts Records

Figure 154. JCL and Required Control Statements

Explanation: In this application, two comparisons are necessary to identify the

records needed for the list: the Grade field (25,2) has to be compared to the
student’s Reading Score field (27,2) and to the Mathematics Score field (29,2). All
numeric fields on the student records are in packed-decimal (PD) format.
The two-clause INCLUDE statement (see Figure 154) guarantees the selection of
the needed records from the file. The first clause (29,2,LT,25,2) guarantees that
records with Math Scores less than the Grade field are INCLUDED. The second
clause (27,2,LT,25,2) guarantees that records with Reading Scores less than the
Grade field are also INCLUDED. The OR connecting the two clauses guarantees
that if either or both of the scores are less than the Grade field, the record is
selected. Finally, since all the fields are in packed-decimal format (PD),
FORMAT=PD is specified.

3–4 Syncsort MFX Programmer’s Guide

 The sample record shown above will be INCLUDED because the student’s Math
Score (047F) is lower than the Grade level (050F).

Omitting Irrelevant Records

Example: Records that have an Invoice Status Code of F (fully paid) are to be
omitted in preparing a list of only those customers with outstanding payments.
(The input record layout is given in Figure 155 and a sample input record is given
in Figure 156.)

Figure 155. Input Record Layout

Figure 156. Sample Input Record

To produce this list of customers selected from the masterfile, the following is
coded.

Syncsort MFX Programmer’s Guide 3–5

 //OUTPAY JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages to
//* I/O Device
//SORTIN DD DSN=NEWINV,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,20) Defines Intermediate Storage
//SYSIN DD *
OMIT COND=(80,1,CH,EQ,C'F') Omits Records
SORT FIELDS=(1,29,CH,A) Sorts Records

Figure 157. JCL and Required Control Statements

Explanation: In this application, a simple comparison is necessary to identify those

masterfile records that are not needed: the Invoice Status Code field (80,1,CH) has
to be compared to the constant 'F'.
The OMIT statement’s condition, 80,1,CH,EQ,C'F', (see Figure 157) guarantees
that invoice records, like the sample record shown above, with the Invoice Status
Code 'F' are omitted from the sort.

Selecting Relevant Fields from the Input Records

Input records often contain some information that is not relevant to a specific
application. For example, records in a personnel masterfile might, in addition to
addresses, include salaries and other confidential information that is not required
for preparing a mailing list.
MFX’s Data Utility features allow you to select only those record fields that contain
necessary data and to eliminate those that do not. More important, MFX enables
you to do this editing before the records are sorted. As a result, the sort has fewer
bytes to handle and processing is more efficient.
For complete syntax of the INREC control statement, see “INREC Control
Statement” on page 2-54.

INREC FIELDS=(p1,l1[,p2,l2,...,pn,ln])

Figure 158. Basic INREC Statement Format

p,l Specify the beginning position and length in bytes of the input
record’s relevant fields. When specifying contiguous fields, or fields
that directly follow one another, you can simply indicate the starting
position of the first field together with the combined length of the
fields that are contiguous.

3–6 Syncsort MFX Programmer’s Guide

Selecting a Number of Fields from Longer Records
Example: A school wants to rank the entire student body by grade point index. This
application simply requires selecting the two relevant fields out of all the fields in
the student records and, then, sorting on the Grade Point Index field. (The Input
Record layout is given in Figure 159.)

Figure 159. Input Record Layout

To include only the relevant fields and generate the ranked list of students, the
following is coded:

//RANK JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=TOT.STUDENTS,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,10),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
INREC FIELDS=(1,9, Selects Record Fields
74,2)
SORT FIELDS=(10,2,PD,D) Sorts Records

Figure 160. JCL and Required Control Statements

Figure 161 shows the input record after INREC processing.

Syncsort MFX Programmer’s Guide 3–7

Figure 161. Form of Post-INREC Record

Explanation: Specifying the two relevant data fields--the Social Security Number
(1,9) and the Grade Point Index (74,2)--on the INREC statement provides the sort
with necessary data for the application and eliminates the fields that are not
relevant to the application. INREC processing thus shortens each record to just a
little under 14% of its original size.

Eliminating Irrelevant Data Field(s)

Example: For an inventory list, the price code on the masterfile records is not
necessary. (The masterfile record layout is given in Figure 162.)

Figure 162. INPUT Record Layout

To eliminate the Price Code field and generate the inventory list, the following is
coded.

3–8 Syncsort MFX Programmer’s Guide

 //INVENTR JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=INV.WARHOUS,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,15),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
INREC FIELDS=(1,17, Selects Record Fields
19,3)
SORT FIELDS=(1,5,CH,A) Sorts Records
.
.
.

Figure 163. JCL and Required Control Statements

Figure 164 shows the input record after INREC processing.

Figure 164. Post-INREC Record Layout

Explanation: Specifying only those fields that are necessary eliminates those that
are not necessary for the application. The Price Code field (18,1) has not been
specified on the INREC statement; it will be deleted from the input records before
the records are sorted by item number for the list.

Selecting Fields from Variable-Length Records

Example: For each volume in its collection, a library requires the catalog number
and any information concerning translations, other volumes in a series, additional
copies on file, and so on. The catalog file consists of variable-length records, and
except for the catalog number, the required information is contained in the
variable-length portion of each record. (The record layout is given in Figure 165.)

Syncsort MFX Programmer’s Guide 3–9

Figure 165. Sample Record Layout

To include only the relevant fields on the input records and to generate this list, the
following is coded.

//LISTCAT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=LIB.CATALOG,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,10),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
INREC FIELDS=(1,14, Selects Record Fields
98)
SORT FIELDS=(5,10,ZD,A) Sorts Records
.
.
.

Figure 166. JCL and Required Control Statements

3–10 Syncsort MFX Programmer’s Guide

 Figure 167 shows the input record after INREC processing.

Figure 167. Form of Post-INREC Record

Explanation: When selecting fields on variable-length records, you must observe

these two restrictions: (1) The position of the RDW cannot be affected; and (2) at
least one byte from the fixed-length portion of the record, in addition to the RDW,
must be specified. On the above INREC statement, the first 14 bytes of each record
– the 4-byte RDW and the fixed-length Catalog Number field – are retained
unchanged. The next field – which contains more information, as required – is
indicated only by position (98) since it is of variable-length. This causes the entire
variable-length portion of the record (beginning with byte 98) to be included after
the initial 14 bytes of the post-INREC record. MFX automatically adjusts the RDW
to reflect the new record length.

Combining Records within a File

Sometimes you may want to shorten a file by consolidating records that have some
information in common. For example, a company’s invoice file may contain more
than one record for any customer to whom multiple invoices have been issued. In
some applications it might then be feasible to consolidate such records – that is, to
combine records with identical Customer Name and Address fields into a single
record containing the sum of that customer’s charges and payments.
The SUM control statement allows you to combine records in this way. For SUM
control statement syntax, see “SUM Control Statement” on page 2-254.

Syncsort MFX Programmer’s Guide 3–11

Combining Records and Summing Numeric Data Fields
Example: For an inventory list, a company requires a single record for each
product, indicating its item number, warehouse code, and the total quantity in
stock. (Figure 168 gives the sample record layout.)

Figure 168. Input Record Layout

To combine those inventory records with identical item numbers and warehouse
codes and to produce the required list, the following is coded.

//INVENT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=WRHSE.INVENT,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,6),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(6,1,CH,A,1,5,ZD,A) Sorts Records
SUM FIELDS=(7,12,PD) Combines Records and
Sums Numeric Data

Figure 169. JCL and Required Control Statements

Explanation: The list is generated by sorting on the Warehouse Code field (6,1,CH)
and the Item Number field (1,5,ZD). Records that have identical information in
both these fields are combined into a single record that contains the sum or total of
those records’ Quantity fields (7,12,PD). That is, the single record will show how
many items with the same number are in each warehouse.

3–12 Syncsort MFX Programmer’s Guide

Eliminating Duplicate Records
Example: A mailing list is being prepared from an invoice file. To eliminate
duplicate entries, any multiple invoice records for the same customer are combined
into a single record. (Figure 170 gives the sample record layout.)

Figure 170. Input Record Layout

To combine multiple invoice records and generate the mailing list, the following is
coded.

//MAILLIST JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=INV.MAST,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
INREC FIELDS=(17,28) Selects Relevant Fields
SORT FIELDS=(1,23,CH,A) Sorts Records. Reference is to
Post-INREC Record
SUM FIELDS=NONE Eliminates Duplicate Records

Figure 171. JCL and Required Control Statements

Explanation: To prepare the customer mailing list, the only information required
from the invoice records is located in the Company Name field (17,23) and the
Address field (40,5), which are selected by the INREC statement. Sorting these
records in ascending order by company name generates an alphabetical list. Then,
because the file contains a record for every transaction, the SUM statement is used
to avoid duplicate listings of customers who have had more than one transaction.
Note that because none of the fields contains numeric data to be summed, the
FIELDS=NONE parameter is used.

Syncsort MFX Programmer’s Guide 3–13

Joining Records from Multiple Files
Sometimes you may want to join two or more files to combine their information for
reports or other purposes. For example, a bank may want to create a report based
on information in three separate files. In some applications it might then be
feasible to join these files – that is, to join the first two files into one, and then
combine that one with a third file for final reporting.

Joining Records
Example: A bank wants to join three separate files to produce a report that shows
recent transactions by customers, sorted by outstanding balance. The final report
shows transaction information from a transaction file, customer name and address
data from a master file containing basic customer information, and the outstanding
balance from a third file containing such information. The record layout for the
transaction file is contained below in Figure 172.

R T
N
BE U N R
M O O BE
U M I
N A T M
N N C U
O O SA R N
TI TI N
A ME
C C R
SA SA T TO
A
N
A
N TE S
A U
TR TR D C
1 78 15 16 26 27
CH CH CH ZD

1 31

Figure 172. Input Record Layout for First File

3–14 Syncsort MFX Programmer’s Guide

 The master file record layout is shown below in Figure 173.

ER S
M E ES
O M R
S T A D
U R N D
C BE A
ER ER
T ER UM M M
S N O
A ST S TO
M U U
C C
1 67 20 21
ZD CH CH

1 22

Figure 173. Input Record Layout for Second File

These two files can be joined on the transaction customer number from the first file
and the master customer number from the second file. These numbers are in zoned
decimal (ZD) format, but because all of the zones are the same in every record,
these fields can be used as character data for the join function.
Figure 174 contains the JCL and control statements to join these two files.

Syncsort MFX Programmer’s Guide 3–15

 //*
//* FIRST STEP
//* JOIN TRANSACTION FILE WITH MASTER FILE CUSTOMER INFO INTO
//* INTERMEDIATE FILE.
//* ADDITIONALLY AN OUTFIL DATA SET WILL BE PRODUCED TO DISPLAY
//* THE JOINED RECORDS CREATED IN THE INTERMEDIATE FILE.
//SORT1 EXEC PGM=SYNCSORT
//* TRANSACTION FILE
//*TRAN# TRANAMT DATE TCUSTNO
//SORTJNF1 DD *
000001 0310.00 12/01/2002 2178I
000002 8055.22 12/02/2002 2123D
000003 0310.00 12/05/2002 2178I
000004 0020.00 12/06/2002 2111A
//*
//* MASTER RECORD FILE
//*MCUSTNO CUSTNAME CUSTADDRESS
//SORTJNF2 DD *
7654C JOSEPH SMITH NY
2111A JAMES JONES NJ
2178I JOHN JACKSON DE
2123D MARY LEE FL
//SORTOUT DD DSN=&&TEMP,DISP=(,PASS),UNIT=SYSDA
//SORTOF01 DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSIN DD *
*
* JOINKEYS FILE=F1,FIELDS=(TCUSTNO)
* JOINKEYS FILE=F2,FIELDS=(MCUSTNO)
* REFORMAT FIELDS=(F1:DATE,TRAN#,TRANAMT,TCUSTNO,F2:CUSTNAME,
* CUSTADDRESS)
*
JOINKEYS FILE=F1,FIELDS=(27,5,ZD,A)
JOINKEYS FILE=F2,FIELDS=(1,5,ZD,A)
REFORMAT FIELDS=(F1:16,11,1,7,8,8,27,6,F2:7,14,21,3)
SORT FIELDS=COPY
OUTFIL FILES=01,HEADER2=('DATE ','TRAN# ','TRANAMT ',
'CUST# ','CUSTOMER NAME ','ADD')
//*

Figure 174. JCL and Required Control Statements

3–16 Syncsort MFX Programmer’s Guide

 Next, a third file containing outstanding balances is added.

E
C
ER N
M LA
O
ST R BA
U G
C BE IN
R
E UM D
N
ST N TA
A
M TS
U
O
1 67
ZD CH

1 15

Figure 175. Input Record Layout for Third File

Syncsort MFX Programmer’s Guide 3–17

 To do that, the following is coded.

//**********************************************************
//*SECOND STEP:
//*JOIN INTERMEDIATE OUTPUT WITH OUTSTANDING BALANCE FILE FOR FINAL
//* REPORT.
//* THE REPORT WILL PRESENT THE DATA WITH HEADINGS INDICATING
//* THE FIELDS PROVIDED.
//*
//SORT2 EXEC PGM=SYNCSORT
//* INTERMEDIATE OUTPUT FILE (FIELDS IN DIFFERENT LOCATION, SO NEED
//* RENAMING)
//*DATE_TEMP,TRAN#_TEMP,TRANAMT_TEMP,TCUSTNO_TEMP,CUSTNAME_TEMP,
//*CUSTADDRESS_TEMP
//SORTJNF1 DD DSN=&&TEMP,DISP=(OLD,DELETE)
//*
//* OUTSTANDING BALANCE FILE
//*MCUSTNO OUTSTANDINGBALANCE
//SORTJNF2 DD *
7654C 00000.00
2111A 09876.54
2178I 00100.00
2123D 13555.22
//SORTOUT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSIN DD *
*
*JOINKEYS FILE=F1,FIELDS=(TCUSTNO_TEMP)
*JOINKEYS FILE=F2,FIELDS=(MCUSTNO)
*REFORMAT FIELDS=(F1:DATE_TEMP,TRAN#_TEMP,TRANAMT_TEMP,TCUSTNO_TEMP,
* CUSTNAME_TEMP,CUSTADDRESS_TEMP,F2:OUTSTANDINGBALANCE)
*SORT FIELDS=(OUTSTANDINGBALANCE_RELOCATED)
*OUTREC FIELDS=(DATE_TEMP_RELOCATED,3X,OUTSTANDINGBALANCE_RELOCATED,3X,
* TRAN#_TEMP_RELOCATED,3X,TRANAMT_TEMP_RELOCATED,3X,
* TCUSTNO_TEMP_RELOCATED,3X,CUSTNAME_TEMP_RELOCATED,3X,
* CUSTADDRESS_TEMP_RELOCATED)
*
JOINKEYS FILE=F1,FIELDS=(27,5,ZD,A)
JOINKEYS FILE=F2,FIELDS=(1,5,ZD,A)
REFORMAT FIELDS=(F1:1,49,F2:7,8)
SORT FIELDS=(50,8,CH,A)
OUTREC FIELDS=(1,11,3X,50,8,3X,12,7,3X,19,8,3X,27,6,3X,33,14,3X,47,3)
OUTFIL FILES=OUT,HEADER2=('DATE ',3X,'OUTSTBAL',3X,
'TRAN# ',3X,'TRANAMT ',3X,
'CUST# ',3X,'CUSTOMER NAME ',3X,'ADD')

Figure 176. JCL and Required Control Statements

3–18 Syncsort MFX Programmer’s Guide

 The output from the first step:

DATE TRAN# TRANAMT CUST# CUSTOMER NAME ADD

12/06/2002 000004 0020.00 2111A JAMES JONES NJ
12/02/2002 000002 8055.22 2123D MARY LEE FL
12/05/2002 000003 0310.00 2178I JOHN JACKSON DE
12/01/2002 000001 0310.00 2178I JOHN JACKSON DE

Figure 177. Sample Output

The output from the second step:

DATE OUTSTBAL TRAN# TRANAMT CUST# CUSTOMER NAME ADD

12/05/2002 00100.00 000003 0310.00 2178I JOHN JACKSON DE
12/01/2002 00100.00 000001 0310.00 2178I JOHN JACKSON DE
12/06/2002 09876.54 000004 0020.00 2111A JAMES JONES NJ
12/02/2002 13555.22 000002 8055.22 2123D MARY LEE FL

Figure 178. Sample Output

Retaining Unpaired Records from One of the Join Files

Example: A bank wants to produce a report of inactive customer accounts. These
are accounts for which there have been no recent transactions. The record layout
for the transaction file was described previously in Figure 96 and the record layout
was described in Figure 97.
This can be done by doing the same join as detailed in Figure 100, but only
retaining the unpaired records from the master file (SORTJNF2) through the use
of the JOIN control statement. This is known as a “right outer join.”

Syncsort MFX Programmer’s Guide 3–19

 //*
//* PRODUCE A REPORT OF INACTIVE CUSTOMERS (THOSE WITH NO TRANSACTIONS)
//* FROM A MASTER FILE WITH CUSTOMER INFO, SORTED BY CUSTOMER NAME.
//*
//SORT1 EXEC PGM=SORT
//* TRANSACTION FILE
//*TRAN# TRANAMT DATE TCUSTNO
//SORTJNF1 DD *
000001 0310.00 12/01/2002 2178I
000002 8055.22 12/02/2002 2123D
000003 0310.00 12/05/2002 2178I
000004 0020.00 12/06/2002 2111A
000005 0033.00 12/06/2002 7654B
000006 1225.00 12/06/2002 2166F
//*
//* MASTER RECORD FILE
//*MCUSTNO CUSTNAME CUSTADDRESS
//SORTJNF2 DD *
7654C JOSEPH SMITH NY
2111A JAMES JONES NJ
2178I JOHN JACKSON DE
2123D MARY LEE FL
0822I MICHAEL JAY CA
//SORTOUT DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSIN DD *
*
* JOINKEYS FILE=F1,FIELDS=(TCUSTNO)
* JOINKEYS FILE=F2,FIELDS=(MCUSTNO)
* JOIN UNPAIRED,F2,ONLY
* REFORMAT FIELDS=(F2:MCUSTNO,CUSTNAME,CUSTADDRESS)
* SORT FIELDS=(CUSTNAME,A)
*
JOINKEYS FILE=F1,FIELDS=(27,5,ZD,A)
JOINKEYS FILE=F2,FIELDS=(1,5,ZD,A)
JOIN UNPAIRED,F2,ONLY
REFORMAT FIELDS=(F2:1,6,7,14,21,3)
SORT FIELDS=(7,13,CH,A)
OUTFIL HEADER2=('INACTIVE CUSTOMERS',2/,
'CUST# ','CUSTOMER NAME ','ADD')
//*

Figure 179. JCL and Required Control Statements

3–20 Syncsort MFX Programmer’s Guide

 The output from this example is:

INACTIVE CUSTOMERS
CUST# CUSTOMER NAME ADD
7654C JOSEPH SMITH NY
0822I MICHAEL JAY CA

Figure 180. Sample Output

Retaining Unpaired Records from Both Join Files

Example: In addition to the Inactive Customers report in the previous example, the
bank can produce an exception report of all transactions for which there is no
master file customer record, all within the same MFX execution. This is done by
also including the unpaired records from the transaction file (a “full outer join”).
The output from the join operation consists of records that either have data from
the unpaired records in the transaction file or the data from the records in the
master file. The missing data will be blanks in each record. These records can then
be reformatted and directed into two separate output file reports, as follows.

Syncsort MFX Programmer’s Guide 3–21

 //*
//* PRODUCE A REPORT OF INACTIVE CUSTOMERS (THOSE WITH NO TRANSACTIONS)
//* FROM A MASTER FILE WITH CUSTOMER INFO, SORTED BY CUSTOMER NAME.
//* ALSO SIMULTANEOUSLY PRODUCE AN EXCEPTION REPORT OF ANY TRANSACTIONS
//* WHERE THE CUSTOMER NUMBER IS UNKNOWN (NO MATCHES IN THE MASTER
//* FILE), SORTED BY DESCENDING TRANSACTION AMOUNT.
//*
//SORT2 EXEC PGM=SORT
//* TRANSACTION FILE
//*TRAN# TRANAMT DATE TCUSTNO
//SORTJNF1 DD *
000001 0310.00 12/01/2002 2178I
000002 8055.22 12/02/2002 2123D
000003 0310.00 12/05/2002 2178I
000004 0020.00 12/06/2002 2111A
000005 0033.00 12/06/2002 7654B
000006 1225.00 12/06/2002 2166F
//*
//* MASTER RECORD FILE
//*MCUSTNO CUSTNAME CUSTADDRESS
//SORTJNF2 DD *
7654C JOSEPH SMITH NY
2111A JAMES JONES NJ
2178I JOHN JACKSON DE
2123D MARY LEE FL
0822I MICHAEL JAY CA
//SORTOF1 DD SYSOUT=*
//SORTOF2 DD SYSOUT=*
//SYSOUT DD SYSOUT=*
//SYSIN DD *
*
* JOINKEYS FILE=F1,FIELDS=(TCUSTNO)
* JOINKEYS FILE=F2,FIELDS=(MCUSTNO)
* JOIN UNPAIRED,ONLY
* REFORMAT FIELDS=(F1:TRAN#,TRANAMT,DATE,TCUSTNO,
* F2:MCUSTNO,CUSTNAME,CUSTADDRESS)
* SORT FIELDS=(CUSTNAME,A,TRANAMT,D)
* OUTFIL FILES=1,INCLUDE=(CUSTNAME,NE,C' '), INACTIVE CUSTOMERS REPT
* HEADER2=(....),
* OUTREC=(CUSTNAME,3X,CUSTADDRESS,3X,MCUSTNO)
* OUTFIL FILES=2,INCLUDE=(TRAN#,NE,C' '), TRANSACTION EXCEPTIONS REPT
* HEADER2=(....),
* OUTREC=(TRAN#,3X,TRANAMT,3X,3X,DATE,3X,TCUSTNO)
*
JOINKEYS FILE=F1,FIELDS=(27,5,ZD,A)
JOINKEYS FILE=F2,FIELDS=(1,5,ZD,A)
JOIN UNPAIRED,ONLY
REFORMAT FIELDS=(F1:1,7,8,8,16,11,27,6,
F2:1,6,7,14,21,3)

Figure 181. JCL and Control Statements (Page 1 of 2)

3–22 Syncsort MFX Programmer’s Guide

 SORT FIELDS=(39,13,CH,A,8,7,CH,D)
OUTFIL FILES=1,INCLUDE=(39,13,CH,NE,C' '), INACTIVE CUSTOMERS REPT
HEADER2=('INACTIVE CUSTOMERS',2/,
'CUSTOMER NAME ',3X,'ADD',3X,'CUST# '),
OUTREC=(39,14,3X,53,3,3X,33,6)
OUTFIL FILES=2,INCLUDE=(1,6,CH,NE,C' '), TRANSACTION EXCEPTIONS REPT
HEADER2=('TRANSACTION EXCEPTIONS',2/,
'TRAN# ',3X,'TRANAMT ',3X,
'DATE ',3X,'CUST# '),
OUTREC=(1,7,3X,8,8,3X,16,11,3X,27,6)
//

Figure 181. JCL and Control Statements (Page 2 of 2)

The SORTOF1 output from this example is:

INACTIVE CUSTOMERS
CUSTOMER NAME ADD CUST#
JOSEPH SMITH NY 7654C
MICHAEL JAY CA 0822I

Figure 182. Sample Output

The SORTOF2 output is:

TRANSACTION EXCEPTIONS
TRAN# TRANAMT DATE CUST#
000006 1225.00 12/06/2002 2166F
000005 0033.00 12/06/2002 7654B

Figure 183. Sample Output

Using Join Processing To Copy a Large Number of Master File

Records
Example: File 1 is a master file with an LRECL of 3400 and contains 140 million
fixed-length records. There is a unique 15-byte printable account number that
appears in column 22 and you want to copy a large number of these records to a
separate output file for further processing. All records in this file are in account
number sequence.
File 2 has an LRECL of 15 and contains 285,000 fixed-length records. The only field
in these records is a 15-byte printable account number. This file is also in account
number sequence.
Using join processing, you can execute an application where File 1 is the master file
and File 2 is your “finder” file. Join processing will create a new output file that will
“copy” only those records with account numbers that exist in File 2.

Syncsort MFX Programmer’s Guide 3–23

 Use the following JCL and control statements:

//JOIN EXEC PGM=SYNCSORT

//SORTJNF1 DD DSN=YOUR.INPUT.MASTER.FILE
//SORTJNF2 DD DSN=YOUR.FINDER.FILE
//SORTOUT DD DSN=YOUR.DESIRED.OUTPUT.FILE
//SYSOUT DD SYSOUT=*
//SYSIN DD *
JOINKEYS FILE=F1,FIELDS=(22,15,CH,A),SORTED
JOINKEYS FILE=F2,FIELDS=(1,15,CH,A),SORTED
REFORMAT FIELDS=(F1:1,3400)
SORT FIELDS=COPY
END
/*

Figure 184. JCL and Control Statements

Following are some sample segments from the input file:

SDKFJSSDFOIEWLKQQQQQQ111111111111111QQQQQQDKDFGSDLFJASDLKFSAKLDLK...
DLKFJASDFOILKLKQQQQQQ222222222222222QQQQQQUSDFSADFASDFSADFSALEOWI...
QDFLKSDKFOIWZWKQQQQQQ333333333333333QQQQQQUCVXCGFLKSFDMXVKLSFGSNK...
SDJKEWUXCVEIITKQQQQQQ444444444444444QQQQQQTUMNSFDSFUEWMNCXVEJRTWW...
JEIVBECKCVWASDKQQQQQQ555555555555555QQQQQQTUMNSFDSFUEWMNCXVEJRTWW...

Figure 185. Master File Excerpts

Following are records from the finder file:

222222222222222
444444444444444

Figure 186. Finder File

Following are segments of the output file:

DLKFJASDFOILKLKQQQQQQ222222222222222QQQQQQUSDFSADFASDFSADFSALEOWI...
SDJKEWUXCVEIITKQQQQQQ444444444444444QQQQQQTUMNSFDSFUEWMNCXVEJRTWW...

Figure 187. Output File Excerpts

Note that the “SORTED” parameter has been added to each JOINKEYS control
statement, because the files are already in the desired sequence. If either file were
not in account number sequence, the application would terminate with an error
message. To address this problem, you would have to remove the applicable
“SORTED” parameter from the corresponding JOINKEYS control statement.

3–24 Syncsort MFX Programmer’s Guide

 Example: This is a variation of the previous example. If you wanted to copy all of
the records in the master file except those with matching account numbers in the
second file, just add the following control statement to the JCL for that example:

JOIN UNPAIRED,F1,ONLY

Figure 188. JOIN Control Statement

This directs join processing to include only the records from File 1, the master file,
that do not have a record in File 2 with the same account number.

Making Output Records Printable and Easy to Read

Because data is usually stored in a compact format, it can be difficult, if not
impossible, to read when printed. For example, on a typical input record, there will
be no blank space between fields, numeric data will sometimes be lost in leading
and trailing zeros, and some data will be in unprintable format.
After processing, you will probably want to edit this data so that it is easy to read.
This is bound to entail one or more of the following tasks:
• reordering the position of record fields
• inserting blanks between fields
• inserting binary zeros
• converting numeric data from unprintable to printable format
• converting data to printable hexadecimal format
• using masks or edit patterns to insert dollar signs, decimal points, slashes, and
the like
• formatting the data in a record field on multiple output lines
MFX’s OUTREC processing, specified either as a control statement or as a
parameter on the OUTFIL statement, can perform these and other editing
functions. The OUTREC control statement is described below. Any number of the
OUTREC statement’s subparameters may be specified and must be coded in the
order in which the fields will appear in the reformatted record. (Note that when
specified as a parameter of OUTFIL, OUTREC is coded identically as for a control
statement except that the keyword FIELDS is not used.) See “OUTREC Control
Statement Format” on page 2-135 for the complete format of the OUTREC
statement.

Syncsort MFX Programmer’s Guide 3–25

Reordering the Positions of Record Fields
Example: A data center has decided to reorder the positions of the data fields in
masterfile records after sorting them. (Figure 189 gives the layout for the
masterfile record.)

Figure 189. Input Record Layout

To sort the records alphabetically by product name and reposition the data fields,
the following is coded:

//SORTPROD JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=PROD.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,10),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(7,15,CH,A) Sorts Records
OUTREC FIELDS=(22,3, Repositions Fields on
7,15, Output Records
1,2,
25,4,
3,4)

Figure 190. JCL and Required Control Statements

Figure 191 shows the output record after OUTREC processing.

3–26 Syncsort MFX Programmer’s Guide

Figure 191. Post-OUTREC Record Layout

Explanation: After the records are sorted alphabetically by product name

(7,15,CH), OUTREC processing moves the Product Code field (22,3) to the first byte
of the record, the Product Name field (7,15) to the fourth byte, the Region field (1,2)
to the nineteenth byte, the Month’s Sales field (25,4) to the twenty-first byte, and
the Sales to Data field (3,4) to the twenty-fifth byte.

Inserting Blanks and Repositioning Record Fields

Example: The central office of a commercial bank requires that each branch
present its masterfile at the end of every month in the format outlined in Figure
192. Branch A, however, has formatted its masterfile records as outlined in Figure
193.

Figure 192. Desired Input Record Layout

Syncsort MFX Programmer’s Guide 3–27

Figure 193. Non-compliant Input Record Layout

To reformat its masterfile records to conform to central office specifications, a bank

branch codes the following. Since the records do not require sorting, the MFX copy
feature is used.

//FORMAT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=ACCT.MAST,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SYSIN DD *
SORT FIELDS=COPY Copies Records
OUTREC FIELDS=(1,4, Repositions Fields on
8,10, Output Records
6X,
5,3,
1X,
18,17)

Figure 194. JCL and Required Control Statements

Figure 195 shows the effect of OUTREC processing on the output record.

3–28 Syncsort MFX Programmer’s Guide

Figure 195. Post-OUTREC Record Layout

Explanation: After the records are copied, OUTREC specifies two types of
reformatting: (1) repositioning data fields and (2) inserting blanks between fields.
As shown in Figure 195, two fields have been repositioned: the Account Type field
now begins on the twenty-first byte as opposed to the fifth byte, and the Account
Number field begins on the fifth byte rather than on the eighth. Also, blanks have
been inserted using the nX entry to specify the number (n) of blanks. Six blanks
have been inserted after the Account Number field and a single blank after the
Account Type field. Since the Balance field and Interest field are contiguous, they
are treated as a single field in this application.

Inserting Binary Zeros

Example: A manufacturing firm has decided to expand its product line. However,
because the Item Number field on its inventory records is too small, the records
must be reformatted to allow for more columns for the new products. The Item
Number is kept in packed-decimal (PD) format, and the firm wants to add 4 bytes
to the current 2 byte field. The new bytes are to precede the current two bytes.
Figure 196 gives the input record layout.

Syncsort MFX Programmer’s Guide 3–29

Figure 196. Input Record Layout

To copy the records and insert the 4 bytes of binary zeros, the following is coded.

//SORTCP JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX
//* Messages to I/O Device
//SORTIN DD DSN=INV.REC,DISP=SHR Defines Input Data Set
//SORTOUT DD DSN=INV.REC.OUT,DISP=(NEW,KEEP), Defines Output Data Set
// UNIT=SYSDA,SPACE=(TRK,5),
// VOL=SER=000111
//SYSIN DD *
SORT FIELDS=COPY Copies Records
OUTREC FIELDS=(1,20, Inserts Binary Zeros &
4Z, Reformats Records
25:21,56)

Figure 197. JCL and Required Control Statements

The effect of OUTREC processing is shown in Figure 198 below.

3–30 Syncsort MFX Programmer’s Guide

Figure 198. Post-OUTREC Record Layout

Explanation: The records are copied, and OUTREC processing adds 4 bytes of
binary zeros (4Z) to the beginning of the Item Number field (21,2). To allow for the
4 additional bytes, the original Item Number field and the fields following it are all
copied after the 4 inserted bytes of zeros.

Converting Unprintable Data to Readable Form

Example: For a file of invoice records sorted by company name, the Invoice Amount,
Amount Paid, and Balance Due fields are to be converted from packed-decimal to
printable format. In addition, any leading zeros will be suppressed and both
commas and decimal points will be inserted. (Figure 199 gives the input record
layout.)

Figure 199. Input Record Layout

Syncsort MFX Programmer’s Guide 3–31

 To sort the records, convert the three fields of packed-decimal data, and insert the
commas and decimal points, the following is coded.

//INVOICE JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX
//* Messages to I/O Device
//SORTIN DD DSN=NEWINV,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate
Storage
//SYSIN DD *
SORT FIELDS=(1,23,CH,A) Sorts Records
OUTREC FIELDS=(17:1,23, Repositions Record Fields
52:24,4,PD,M2, and Converts Data
74:28,4,PD,M2,
96:32,4,PD,M2)

Figure 200. JCL and Required Control Statements

The effect of OUTREC processing on the input record is shown in Figure 201 below.

Figure 201. Post-OUTREC Record Layout

Explanation: First the records are sorted alphabetically by company name

(1,23,CH). Then, three fields--the Invoice Amount (24,4,PD), the Amount Paid
(28,4,PD), and the Balance Due (32,4,PD)--are converted from packed-decimal (PD)
into readable format and editing by an MFX editing mask (M2) that suppresses the
printing of leading zeros and inserts the appropriate commas and decimal points.
The number-colon entries (c:) that precede each of the four fields assign a new
starting position or, when printing, column for each of the four fields. For example,
the Company Name field, which originally began in byte 1 for a length of 23 bytes,
now begins in byte 17; the Invoice Amount field, which began in byte 24, begins in
byte 52, and so on. Note that after the data is converted and edited, the lengths of

3–32 Syncsort MFX Programmer’s Guide

 the packed-decimal fields increase from four bytes each to ten bytes and that the
fields are each separated by twelve blanks.

Converting Unprintable Data to Hexadecimal Format

Example: A bank has discovered that some errors were made in recording the
Account Numbers of some of its customers. Specifically, on the transaction records,
some Account Number fields, which should contain only packed-decimal, PD, data,
appear to contain data that is not valid packed-decimal. Figure 202 shows the
input record layout.

Figure 202. Sample Input Record Layout

In order to find the invalid data, the following is coded.

//SORTHEX JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX
//* Messages to I/O Device
//SORTIN DD DSN=TRANS.RECS,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SYSIN DD *
SORT FIELDS=COPY Copies Records
OUTREC FIELDS=(1,30, Reformats Output Records
36:31,12,HEX) and Converts Data

Figure 203. JCL and Required Control Statements

The effect of OUTREC processing on the input record is shown in Figure 204.

Syncsort MFX Programmer’s Guide 3–33

Figure 204. Sample Post-OUTREC Record Layout

Explanation: The records are copied, and OUTREC processing reformats the
output record to contain the Customer Name field (1,30) followed in column 36 by
the Account Number field converted to hexadecimal format (31,12,HEX). Blanks
are automatically inserted in the unspecified columns (31,5). Note that converting
the Account Number data to printable hexadecimal expands the original 12-byte
field to 24 bytes. The bank can now read the Account Number field in hexadecimal
format to determine which records contain invalid data.

Converting and Editing Unprintable Data

Example: For an Outstanding Payments report, the packed-decimal Amount Due
field on a company’s invoice records is converted to printable format and edited
with a floating dollar sign, commas, and a decimal point. In addition, to make the
output easy to read, ten blanks are inserted between the Company Name field and
the Amount Due field. (Figure 205 gives the input record layout.)

3–34 Syncsort MFX Programmer’s Guide

Figure 205. Input Record Layout

To sort the records and accomplish the conversion and editing, the following is
coded.

//PAYMNT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=INVOICE,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,23,CH,A) Sorts Records
OUTREC FIELDS=(1,23, Converts and Edits Data
10X, and Inserts Blanks
24,4,PD,EDIT=($II,IIT.TT))

Figure 206. JCL and Required Control Statements

Figure 207 shows the effect of OUTREC processing on the input record.

Syncsort MFX Programmer’s Guide 3–35

Figure 207. Post-OUTREC Record Layout

Explanation: First the records are sorted alphabetically by Company Name

(1,23,CH). Next, OUTREC processing inserts 10 blanks (10X) between the
Company Name field (1,23) and the Balance Due field (24,4,PD). OUTREC
processing also converts this packed-decimal field to printable format and edits it
with the user-provided pattern specified on the EDIT subparameter,
EDIT=($II,IIT.TT). This pattern provides for a floating dollar sign as well as the
appropriate comma and decimal point. The Is indicate that leading zeros should not
be printed and the Ts indicate that zeros in those positions should be printed. Note
that this conversion and editing of the data cause the length of the Balance Due
field to increase from its original length of four bytes to ten bytes.

Putting a Data Field in Standard Format

Example: The date field on insurance-policy records is stored in zoned-decimal
format but without slashes separating the month, day, and year. After the records
are sorted, these slashes will be inserted and the date will appear in the standard
mm/dd/yy format. (Figure 208 gives the input record layout.)

3–36 Syncsort MFX Programmer’s Guide

Figure 208. Input Record Layout

To sort the records and format the date field with the required slashes, the
following is coded.

//SORTDT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=NEW.POLCY,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,23,CH,A) Sorts Records
OUTREC FIELDS=(1:1,23, Edits Data and Repositions
30:24,6,ZD,M9, Record Fields
45:30,8)

Figure 209. JCL and Required Control Statements

The effect of OUTREC processing is shown in Figure 210.

Syncsort MFX Programmer’s Guide 3–37

Figure 210. Post-OUTREC Record Layout

Explanation: The records are sorted alphabetically by Member Name (1,23,CH).

The OUTREC statement repositions the Effective Date field (24,6,ZD) and the
Policy Number field (30,8,ZD) in columns 30 and 45 respectively, leaving blanks
between each of the three fields. In addition, the OUTREC statement edits the
Effective Date field with an M9 editing mask that places slashes between the
month, date, and year. Note that editing the Date field increases its size from six to
eight bytes.

Converting from Variable to Fixed-Length Format

Example: In this example, there are three output files. The first is variable and the
remaining two are fixed-length format. The variable output file is the standard
output file from the sort. In order to convert the output from variable to fixed-
length format, you should specify CONVERT on the OUTREC parameters of each
of your OUTFIL control statements. The following are the JCL and control
statements to effect this result.

3–38 Syncsort MFX Programmer’s Guide

 // JOB
// EXEC PGM=SYNCSORT
//SYSOUT DD SYSOUT=*
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA
//SORTIN DD DSN=VARIN,DISP=SHR
//SORTOUT DD UNIT=SYSDA,SPACE=(CYL,(1,1)),
// DISP=(,PASS),DSN=&&VAROUT
//SORTOF1 DD UNIT=SYSDA,SPACE=(CYL,(1,1)),
// DISP=(,PASS),DSN=&&FIX1OUT
//SORTOF2 DD UNIT=SYSDA,SPACE=(CYL,(1,1)),
// DISP=(,PASS),DSN=&&FIX2OUT
//SYSIN DD *
SORT FIELDS=(5,19,CH,A,28,2,CH,A)
OUTFIL FILES=1,
INCLUDE=(28,2,CH,EQ,C'92'),
OUTREC=(5,19),CONVERT
OUTFIL FILES=2,
INCLUDE=(28,2,CH,EQ,C'93'),
OUTREC=(5,19),CONVERT

Figure 211. Using the CONVERT Parameter

Printing Input Records on Multiple Output Lines

Example: In this example, five input record fields, shown in Figure 212, are copied
to an output file with each field printed as a separate output line.

Figure 212. Input Record Layout

Multiple output lines are created by specifying a new line character, i.e. / (slash), in
the OUTREC parameter of an OUTFIL control statement. As shown in Figure 213,
the new line character follows the specification of each input field’s starting
position and length.

Syncsort MFX Programmer’s Guide 3–39

 //MULTILIN JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=&&DATA,DISP=SHR Defines Input Data
//* Set
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(3,3)) Defines Intermediate
//* Storage
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(3,3)) Defines Intermediate
//* Storage
//SORTOUT DD SYSOUT=* Defines Output Data
//* Set
//SYSIN DD *
SORT FIELDS=(101,40,CH,A) Sorts Records
OUTFIL CONVERT, Converts Data
HEADER2=('CUSTOMER ADDRESS LIST',3/), Prints a Page
* Header
OUTREC=(101,40,/, Prints the Data in the
* Field and Starts a
* New Output Line
141,25,/, As Above
166,25,/, As Above
191,30,/, As Above
266,35,2/) As Above but Starts
* 2 New Output Lines

Figure 213. JCL and Control Statements for Multiline Output

Once MFX has printed the data in the COMPANY NAME field, it starts a new
output line, prints on it the data in the next field, CUSTOMER NAME, starts a
new line, and so forth. After printing the contents of the last field (CITY, STATE
AND ZIP), MFX creates two new lines (2/).
Figure 214 provides an excerpt from the output file where the input record is
formatted on multiple lines. A blank line appears in the second and third set of
multiline output because the corresponding input record fields (i.e. CUSTOMER
TITLE and CUSTOMER NAME) were blank.

3–40 Syncsort MFX Programmer’s Guide

 CUSTOMER ADDRESS LIST

AARON'S ROD INC. First Set of Multiline Output

DAVID LAURENCE
SYS PROG
6936 YOUNGMAN BLVD.
GREAT NECK CT 06854

BLAKE'S VISION TECHNOLOGY Second Set of Multiline Output

MR. N. FRYE

261 ALBION PLACE

SEA BRIGHT NJ 08572

COLTRANE & COMPANY Third Set of Multiline Output

DATA CENTER MANAGER

300 DORIAN AVENUE
NEW YORK NY 11220

Figure 214. Sample Multiline Output

Dividing a Report into Sections

When printing sorted output, you may want to divide it into sections. For example,
after sorting a personnel file alphabetically by company name and department, you
might want to print each department’s records as a separate section and leave
some blank lines between each section. You might even want to print each section
as a separate page of the report. MFX allows you to print groups of records that
have identical information in one or more sort fields as sections and to separate
each section by a specified number of lines or a page break.
To divide output into sections, use the SECTIONS parameter on the OUTFIL
control statement. For complete syntax of the SECTIONS parameter, see
“SECTIONS Parameter (Optional)” on page 2-118.

Dividing Output into Sections

Example: A personnel roster is to be divided into sections by Department. (Figure
215 presents the layout for the input record.)

Syncsort MFX Programmer’s Guide 3–41

Figure 215. Input Record Layout

To sort the records and generate a list that is divided by Department, the following
is coded.

//ROSTER JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=PRSNL,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,2),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(15,5,A,1,14,A),FORMAT=CH Sorts Records
OUTFIL OUTREC=(6:15,5, Repositions Record Fields
14:1,14,
33:20,3,
44:23,1,
54:24,2),
SECTIONS=(15,5,SKIP=5L) Sections Records

Figure 216. JCL and Required Control Statements

A sample of the listing generated is shown in Figure 217.

3–42 Syncsort MFX Programmer’s Guide

 ACCTG BELL PAT SUP F 03
ACCTG EMERY PAUL CLK M 04
ACCTG JONES MARK CLK M 01
ACCTG NORTH NANCY MGR F 02
ACCTG OWEN JERRY CLK M 03
ACCTG TWAIN JOAN SEC F 05
ACCTG WEST DONNA CLK F 03

PRSNL SMITHE JON CLK M 00

PRSNL TOWERS LINDA CLK F 02
PRSNL VREES GEORGE CLK M 02
PRSNL WU JANE SUP F 05
PRSNL YOUNG RUSS MGR M 03

Figure 217. Sample Output

Explanation: After the records are sorted alphabetically by Department (15,5) and
Employee Name (1,14), they are divided into sections by department. That is, every
time there is a change in the Department field (15,5 in the input record) the printer
skips 5 lines (5L) before printing the next record. (Note, in the Sample Output
above, the five-line break that occurs between ACCTG and PRSNL.) The OUTREC
parameter is used to reposition the record fields and to leave blanks between them.

Writing Headers and Trailers for a Report

Headers are used to provide report, page, and section headings such as titles, page
numbers, the current date, labels for each column of data, and the like. Similarly,
trailers are used for report, page, and section summaries. You can use them, for
example, to provide totals for columns of numeric data (see “Totaling and
Subtotaling Data” on page 3-51) or to indicate the end of a section with, say, a
string of asterisks or to provide a list of abbreviations used in the report.
To generate Headers and/or Trailers, use the HEADER and TRAILER parameters
of the OUTFIL control statement. For complete syntax, see “HEADER1/HEADER2
Parameters (Optional)” on page 2-98 and “TRAILER Parameters (Optional)” on
page 2-105.

Writing a Title Page for a Report

Example: Marketing wants a title page for its monthly departmental sales report.
The three-line title will begin on line 16 and three blank lines will separate each
line of the title. The three lines will start printing in columns 49, 59, and 63,
respectively.
To print this title page, the following is coded:

Syncsort MFX Programmer’s Guide 3–43

 //DSRPT JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
.
.
.
OUTFIL HEADER1=(15/,49:'D E P A R T M E N T A L S A L E S',
4/,59:'F E B R U A R Y',
4/,63:'2 0 0 4'), Generates Title Page
.
.
.

Figure 218. JCL and Required Control Statements

Figure 219 shows the header that is generated by the above HEADER1 parameter:

D E P A R T M E N T A L S A L E S

F E B R U A R Y

2 0 0 4

Figure 219. Sample HEADER1

Explanation: The HEADER1 parameter produces a header that will print on a

separate page, with no page number, at the beginning of the report. The first
number-slash (n/) entry, 15/, causes the printer to skip 15 lines before printing. The
following number-colon entry (c:), 49:, specifies the column in which the literal
string 'D E P A R T M E N T A L S A L E S' begins to print. Note that the literal
string prints exactly as it is entered between the single quotes, with a space between
each letter and a double space between the words.
The next entry, 4/, causes the printer to skip 3 more blank lines before starting to
print the literal string 'F E B R U A R Y' in column 59.
Finally, three more lines are left blank (4/) and the literal string '2 0 0 4' begins
printing in column 63.

Writing a Page Header

Example: Marketing wants the first line of every page of its departmental sales
report to contain the program number, report title, page number, and date. They

3–44 Syncsort MFX Programmer’s Guide

 want the third line of every page to contain an identifying label for each column of
data. Each of these lines will begin printing in column one.
To print the page header, the following is coded.

//DSRPT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
.
.
.
OUTFIL .
.
HEADER2=(1:'PGM NUMBER 5',
46:'DEPARTMENT SALES REPORT FOR FEBRUARY 1992',
101:'DATE:',
107:&DATE, Generates Page Heading
121:'PAGE:',
127:&PAGE,//,
1:'DEPARTMENT',
40:'SALES MANAGER',
61:'SALES REP',
78:'SALES THIS PERIOD',
103:'SALES YEAR TO DATE',//),
.
.

Figure 220. JCL and Required Control Statements

Figure 221 shows a representation of the header that is generated by the above
HEADER2 parameter.

PGM NUMBER 5 DEPARTMENT SALES REPORT FOR FEBRUARY 1992 DATE: 02/01/
92 PAGE: 1

DEPARTMENT SALES MANAGER SALES REP SALES THIS PERIOD SALES YEAR TO DA
TE

Figure 221. Sample HEADER2

Explanation: The HEADER2 parameter produces the page header shown above.
Because no forward spacing is specified, the page header begins on the first line of
every page. Each of the HEADER2’s number-colon entries (c:), for example, 1:,
indicates the column in which the entry following the colon begins to print. Thus,

Syncsort MFX Programmer’s Guide 3–45

 the literal 'PGM NUMBER 5' is printed beginning in column 1, and so on. The
&DATE and the &PAGE entries generate a current date and a consecutive page
number, respectively. The date and the page number appear after the labels DATE:
and PAGE:, which are specified like the other literals.
The double slashes (//) following the &PAGE entry direct the printer to forward
space two lines, that is, to leave one blank line, before printing the next group of
literals that constitute the labels for the columns of data.

Writing a Section Header

Example: Marketing wants each section of its departmental sales report to have its
own heading. The heading will consist of one line containing an identifying label for
each column of data. The heading will begin printing in column one.
To print the section header, the following is coded.

//DSRPT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
OUTFIL OUTREC=(1:1,15, Repositions Fields on Output
23:23,7, Records and Edits Data
51:48,3,
72:60,4,PD,EDIT=($II,IIT.TT),
101:64,4,PD,EDIT=($II,IIT.TT),
114:C' '),
SECTIONS=(1,15,SKIP=5L, Generates Section Breaks
HEADER3=(1:'DEPARTMENT', Generates Section Headings
23:'SALES MGR',
48:'SALES REP',
68:'SALES THIS PERIOD',
97:'SALES YEAR TO DATE',//))

Figure 222. JCL and Required Control Statements

Figure 223 shows the header that is generated by the above HEADER3
subparameter.

3–46 Syncsort MFX Programmer’s Guide

 DEPARTMENT SALES MGR SALES REP SALES THIS PERIOD SALES YEAR TO DATE

OVER COUNTER CASEY 075 $14,000.00 $27,000.00

OVER COUNTER CASEY 093 13,550.00 32,000.00

OVER COUNTER CASEY 084 11,755.00 24,850.00

OVER COUNTER CASEY 090 12,250.00 25,000.00

OVER COUNTER CASEY 095 13,075.00 26,180.00

DEPARTMENT SALES MGR. SALES REP SALES THIS PERIOD SALES YEAR TO DATE

SURGICAL KILDARE 003 $11,750.00 $25,320.00

SURGICAL KILDARE 007 $14,300.00 24,900.00

SURGICAL KILDARE 009 11,110.00 30,850.00

SURGICAL KILDARE 004 13,375.00 27,505.00

. . . . .
. . . . .
. . . . .

Figure 223. Sample Sections with HEADER3

Explanation: The HEADER3 subparameter on the SECTIONS parameter

generates a header that prints at the beginning of each section. Its primary
purpose here is to provide labels for the columns of data that appear in each
section. Each of the number-colon entries (c:) specifies the column in which the
entry following it should begin to print. Thus, the literal string 'DEPARTMENT'
begins to print in column 1, the literal string 'SALES MGR' begins to print in
column 23, and so on. Blanks are automatically inserted in the space between the
columns that are specified. On the OUTREC parameter a blank has been inserted
in column 114 (114:C' ') so that the output record length will equal that of the
header. Note that if the HEADER3 in this example were used in conjunction with
the preceding HEADER2 example, there would be no need to specify the labels for
the columns of data in the HEADER2.

Using a Header to Eliminate Duplicate Information within a Section

Example: Rather than repeat the department name and sales manager, which are
identical for every record included in a section of the departmental sales report,
marketing wants this information to appear only once-within the section headers of
the report. Therefore, the section headers’ first two entries (Department and Sales
Manager) will be drawn directly from the first data record in each section.
To print the section header with the input data fields, the following is coded.

Syncsort MFX Programmer’s Guide 3–47

 //DSRPT JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
.
.
.
OUTFIL..., Repositions Fields on Output
OUTREC=(25:48,3, Records and Edits Data
37:60,4,PD,EDIT=($II,IIT.TT),
56:64,4,PD,EDIT=($II,IIT.TT),
71:C' '),
SECTIONS=(1,15,SKIP=2L, Generates Section Breaks
HEADER3=(1:1,15, Generates Section Headings
16:23,7,
23:'SALES REP',
34:'SALES THIS PERIOD',
54:'SALES YEAR TO DATE'))

Figure 224. JCL and Required Control Statements

Figure 225 shows the header that is generated by the above HEADER3
subparameter.

OVER COUNTER CASEY SALES REP SALES THIS PERIOD SALES YEAR TO DATE
075 $14,000.00 $27,000.00
093 $13,550.00 $32,000.00
084 $11,755.00 $24,850.00
090 $12,250.00 $25,000.00
095 $13,075.00 $26,180.00
SURGICAL KILDARE SALES REP SALES THIS PERIOD SALES YEAR TO DATE
003 $11,750.00 $25,320.00
007 $14,300.00 $24,900.00
009 $11,110.00 $30,850.00
004 $13,375.00 $27,505.00
. . .
. . .
. . .

Figure 225. Sample Sections with HEADER3 Including Data from Input Record

Explanation: The HEADER3 subparameter on the SECTIONS parameter

generates a header that prints at the beginning of each section. Its primary
purpose here is to provide individualized section headings that contain the
Department Name and the Sales Manager from the records in that section as well

3–48 Syncsort MFX Programmer’s Guide

 as labels for the columns of data. The first two entries in this header, 1:1.15 and
16:23,7 (the Department Name and Sales Manager, respectively), are drawn
directly from the input record to eliminate the repetition of these fields in the detail
lines of each section. Note that specifying these fields in the HEADER3 eliminates
the need to include them in OUTREC processing as was necessary in the preceding
example. Each of the number-colon entries (c:) specifies the column in which the
entry following it should begin to print. Thus, the Department field, (1,15) begins to
print in column 1; the Sales Manager field, in column 16; the literal string "SALES
REP", in column 48, and so on. Blanks are automatically inserted in the space
between the columns that are specified. It should be pointed out that on the
OUTREC parameter a blank has been inserted in column 71 (71:C' ') so that the
output record length will equal that of the header.

Writing a Report Trailer or Summary

Example: The final page of marketing’s departmental sales report will contain a
note saying that February sales figures include residual 1992 sales not previously
recorded. This note will begin on the 21st line of the page and start printing in the
33rd column of the page.
To print the report trailer, the following is coded.

//DSRPT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
.
.
.
OUTFIL .
.
.
TRAILER1=(20/, Generates Report Trailer
33:'FEBRUARY SALES FIGURES INCLUDE RESIDUAL 1992',
'SALES NOT PREVIOUSLY RECORDED')

Figure 226. JCL and Required Control Statements

Figure 123 shows the trailer that is generated by the above TRAILER1 parameter.

FEBRUARY SALES FIGURES INCLUDE RESIDUAL 1992 SALES NOT PREVIOUSLY RECORDED

Figure 227. Sample TRAILER1

Syncsort MFX Programmer’s Guide 3–49

 Explanation: The TRAILER1 parameter produces a report trailer or summary that
constitutes the final page of a report. Unless otherwise specified, it begins on the
first line of the page. The TRAILER1’s initial number-slash (n/) entry, 20/, directs
the printer to forward space 20 blank lines before printing on the 21st line. The
next entry, a number-colon (c:) entry, is used to center the literal string that follows
it by having the string of characters begin printing in the appropriate column. It
specifies column 33 as the beginning position for printing the literal string,
'FEBRUARY SALES FIGURES INCLUDE RESIDUAL 1992 SALES NOT
PREVIOUSLY RECORDED'.

Writing a Page Trailer

Example: Marketing wants the last line on every page of its departmental-sales
report to contain a note identifying the information as confidential. This line will
begin printing in column one.
To print the page trailer, the following is coded.

//DSRPT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
.
.
.
OUTFIL .
.
.
TRAILER2=(5'*','C O N F I D E N T I A L I N F O R M A T I O N',
5'*','C O N F I D E N T I A L I N F O R M A T I O N',5'*')
.
. Generates Page Trailer
.

Figure 228. JCL and Required Control Statements

Figure 229 shows the trailer that is generated by the above TRAILER2 parameter.

*****CONFIDENTIAL INFORMATION*****CONFIDENTIAL INFORMATION*****

Figure 229. Sample TRAILER2

Explanation: The TRAILER2 coded above provides a trailer that appears at the
bottom of every logical page. The first entry, 5'*', a literal enclosed in single quotes
(in this case an asterisk) and a repetition factor (5), specifies that 5 asterisks

3–50 Syncsort MFX Programmer’s Guide

 should be printed. Because no column was specified, the trailer begins in column
one. The next entry, 'C O N F I D E N T I A L I N F O R M A T I O N ', specifies
that the literal string enclosed in the single quotes should directly follow the
asterisks. Note that the literal string is printed exactly as it is coded within the
quotation marks. That is, there is a blank between every letter and two blanks
between each word. The trailer’s other entries specify the printing of another five
asterisks followed by the literal string 'C O N F I D E N T I A L I N F O R M A T I
O N ' and finally another five asterisks.

Totaling and Subtotaling Data

Writing a summary or trailer for a report will sometimes involve providing totals
for columns of figures. For example, you would probably want a trailer for an
inventory report to contain the total number of items on hand. The OUTFIL
statement allows you to write trailers that contain both totals and subtotals.
Moreover, you can total data at the end of a report, at the end of a page, and also at
the end of a section.
To generated total and subtotals, use the TOTAL and SUBTOTAL entries of
OUTFIL’s TRAILER parameters and subparameter. For details of syntax, see
“TRAILER Parameters (Optional)” on page 2-105

Totaling Data at the End of a Report

Example: The departmental sales report’s final page will be a summary containing
both the total for the sales this period and the total for the sales to date. The trailer
will begin on the 21st line of the page and each total will have an identifying label.
To print the report trailer, the following is coded.

Syncsort MFX Programmer’s Guide 3–51

 //DSRPT JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=MRKTNG.SALES, Defines Input Data Set
DISP=SHR
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5), Defines Intermediate Storage
UNIT=SYSDA
//SYSIN DD *
SORT FIELDS=(1,15,CH,A) Sorts Records
.
.
.
OUTFIL.
.
.
TRAILER1=(20/, Generates Report Trailer with Totals
40:'SALES THIS PERIOD:',
59:TOT=(24,4,PD,EDIT=($II,IIT.TT)),
73:'SALES TO DATE:',
88:TOT=(28,4,PD,EDIT=($II,IIT.TT)))

Figure 230. JCL and Required Control Statements

Figure 231 shows the trailer that is generated by the above TRAILER1 parameter.

SALES THIS PERIOD: $35,807.85 SALES TO DATE: $62,305.25

Figure 231. Sample TRAILER1

Explanation: The TRAILER1 parameter produces a report trailer or summary that

constitutes the final page of a report. Unless otherwise specified, it begins on the
first line of the page. This TRAILER1’s initial number-slash(n/) entry, 20/, directs
the printer to forward space 20 blank lines before printing. The next entry, a
number-colon (c:) entry, is used to center the literal string that follows it by having
the string of characters begin printing in the appropriate column. It specifies
column 40 as the beginning position for the literal string 'SALES THIS PERIOD:'
that labels the numeric data following it. This TRAILER’s other number-colon plus
literal-string entry functions the same way.
The two TOT entries, TOT=(....), generate the trailer’s totals. These entries specify
the numeric data used and its format. Thus the four bytes of packed-decimal data
that begin in byte 24 (24,4,PD) and the four bytes that begin in byte 28 (28,4,PD) of
the input record are converted to printable format. This data is then edited by the
EDIT pattern ($II,IIT.TT), which suppresses the printing of leading zeros and
inserts a floating dollar sign as well as a necessary comma and decimal point. The
pattern uses an I to indicate those zeros in the total that should not be printed and
a T to indicate those that should.

3–52 Syncsort MFX Programmer’s Guide

 Note: Be sure to code all the necessary parentheses when using the TOTAL and
EDIT entries.

Subtotaling Data at the End of a Page

Example: The page trailer for a report listing invoices is to contain the totals for the
Amount Paid and the Balance Due fields of the invoice records printed up to and
including that page. These totals will appear directly below the columns of figures
and be separated from them by strings of hyphens. An identifying label, TOTALS:,
will appear on the same line as the totals and will begin in column 40.
To generate the trailer, the following is coded.

//INVLST JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=INVOICE,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(9,23,A,36,2,A,32,4,A), Sorts Records
FORMAT=CH
.
.
.
OUTFIL.
.
.
TRAILER2=(65:10'-', 86:10'-',/, Generates Page Trailer
40:'TOTALS:', with Running Totals
65:SUB=(46,4,PD,EDIT=($II,IIT.TT)),
86:SUB=(54,4,PD,EDIT=($II,IIT.TT)))

Figure 232. JCL and Required Control Statements

Figure 233 shows the trailer that is produced.

Syncsort MFX Programmer’s Guide 3–53

 . . . . .
. . . . .
. . . . .
MERLINS TRUST CO 82124054 12/15/92 0.00 1,500.00
MEWER COLLEGE 83013324 1/17/92 0.00 1,500.00
NORTHEAST INDUST 83013303 1/17/92 200.00 200.00
PARK PLACE CORP 83022211 2/15/92 0.00 650.00
PATIO PRODUCTS 83022203 2/15/92 0.00 850.00
PINES ASSOCIATES 83022587 2/15/92 0.00 750.00
POLL DATA CORP 82124019 12/15/92 0.00 600.00
PRIESTLEY METALS 83022201 2/15/92 0.00 1,600.00
REGENCY TRUST CO 82124011 12/15/92 0.00 1,500.00
REPUBLIC DATA 83013306 1/17/92 0.00 1,100.00
RIBBIT TECHNOLOGIES 82124020 12/15/92 0.00 360.00
RICE FEATURES 82124015 12/15/92 750.00 750.00
RICE FEATURES 83013298 1/17/92 0.00 1,500.00
RICE FEATURES 83022198 2/15/92 0.00 1,500.00
ROBINS NEST CORP 83013353 1/17/92 0.00 900.00
SIDNEY COLLEGE 82124016 12/15/92 0.00 5,000.00
SIDNEY COLLEGE 83013297 1/17/92 0.00 2,500.00
------- ----------
TOTALS: $6,150.00 $66,475.00

Figure 233. TRAILER2 with SUBTOTAL

Explanation: The above TRAILER2 provides for totaling the figures in the Amount
Paid field (46,4,PD) and the Amount Due field (54,4,PD) on the invoice records.
Because the SUB (SUBTOTAL) entry is specified, the totals that appear at the
bottom of each page represent running totals, that is, the totals for all the records
that have been printed up to and including that page. The TRAILER2 also
generates the identifying label TOTALS: (40:'TOTALS:') and strings of hyphens at
the bottoms of the columns to be totaled (65:10'-', 86:10'-').
The totaled data for each field is converted to printable format and, after being
edited, begins printing in the columns specified with the two number colon entries
(c:), 65: and 86:. The data is edited by the EDIT pattern, ($II,IIT.TT), which
suppresses the printing of leading zeros and inserts a floating dollar sign as well as
the necessary comma and decimal point. The pattern uses an I to indicate the zeros
in the total that should not be printed and a T to indicate those that should.

Totaling Data at the End of a Section

Example: The section trailer for an accounts receivable report sectioned by month
is to contain the totals for the Amount Paid and the Balance Due columns of each
section. These totals will appear directly below the columns of figures and be
separated from them by strings of hyphens. An identifying label, TOTALS:, will
appear on the same line as the totals and will begin in column 40.
To generate the trailer, the following is coded.

3–54 Syncsort MFX Programmer’s Guide

 //ACTREC JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=NEW.INV,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(9,23,A,36,2,A,32,4,A), Sorts Records
FORMAT=CH
.
.
.
OUTFIL.
.
.
SECTIONS=(32,4,SKIP=3L, Generates Section Breaks
TRAILER3=(65:10'-',86:10'-',/, Generates Section Trailer
40:'TOTALS:', with Totals
65:TOT=(46,4,PD,EDIT=($II,IIT.TT)),
86:TOT=(54,4,PD,EDIT=($II,IIT.TT))))

Figure 234. JCL and Required Control Statements

Figure 235 shows the section trailer, with totals, that is produced.

Syncsort MFX Programmer’s Guide 3–55

 . . . . .
. . . . .
. . . . .
WINIFRED INDUST 82124013 12/15/91 300.00 350.00
--------- ----------
TOTALS: $2,600.00 $19,770.00

ARLINE FRAGRANCES 83013304 1/17/92 0.00 7,500.00

CHARACTER DATA 83013343 1/17/92 0.00 1,100.00
COUNTRY INDUSTRIAL 83013557 1/17/92 0.00 950.00
DUNHAM INDUST INC 83013302 1/17/92 0.00 850.00
ECHO LABS INC 83013300 1/17/92 0.00 550.00
ESS SECURITIES 83013311 1/17/92 0.00 550.00
EVERMORE INDUST 83013556 1/17/92 2,000.00 3,000.00
GOODEY FOODS 83013356 1/17/92 0.00 600.00
GROSS BOOKS CO 83013264 1/17/92 0.00 2,500.00
HARVEY MOTORS CO 83013301 1/17/92 2,000.00 3,000.00
KALABRA CORPORATION 83013555 1/17/92 0.00 1,500.00
MEWER COLLEGE 83013324 1/17/92 0.00 1,500.00
NORTHEAST INDUST 83013303 1/17/92 200.00 200.00
REPUBLIC DATA 83013306 1/17/92 0.00 1,100.00
RICE FEATURES 83013298 1/17/92 0.00 1,500.00
ROBINS NEST CORP 83013353 1/17/92 0.00 900.00
SIDNEY COLLEGE 83013297 1/17/92 0.00 2,500.00
SOUTHWEST INDUST 83013503 1/17/92 200.00 200.00
SPENSERS INDUST 83013989 1/17/92 0.00 650.00
UNITED INTERESTS INC 83013309 1/17/92 0.00 1,500.00
WINIFRED INDUST 83013299 1/17/92 0.00 650.00
---------- ---------
TOTALS: $4,400.00 $32,800.00

Figure 235. TRAILER3 with TOTAL

Explanation: In addition to generating strings of hyphens at the bottom of the

columns to be totaled (65:10'-',86:10'-') and the identifying label TOTALS: on the
line below (40:'TOTALS:'), the TRAILER3 provides for totaling the figures in the
Amount Paid field (46,4,PD) and the Amount Due field (54,4,PD) on the invoice
records. Note that because the TOT (TOTAL) entry is specified, the totals that
appear at the end of each section represent that totals only for the records that are
included in that section.
The totaled data for each field is converted to printable format and, after being
edited, begins printing in the columns specified with the two number colon entries
(c:), 65: and 86:. The data is edited by the EDIT pattern, ($II,IIT.TT), which
suppresses the printing of leading zeros and inserts a floating dollar sign as well as
the necessary comma and decimal point. The pattern uses an I to indicate the zeros
in the total that should not be printed and a T to indicate those that should.

3–56 Syncsort MFX Programmer’s Guide

Obtaining Maximum, Minimum and Average Data
A report may need to include maximum, minimum, and average data. The
parameters provided for this type of reporting are MIN, SUBMIN, MAX, SUBMAX,
AVG and SUBAVG. The syntax is the same as for TOTAL and SUBTOTAL. See
“Totaling and Subtotaling Data” on page 3-51 and “TRAILER Parameters
(Optional)” on page 2-105.

Printing Maximum, Minimum and Average Data in Section Trailers

Example: The section trailers for an accounts receivable report sectioned by data
group (AAA, BBB, etc.) are to contain six edited numeric values for a 6-byte field
that begins at byte 8 (8,6). The values to be printed are the following:
• The minimum data value up to that point in the report (SUBMIN)
• The minimum data value in the section (MIN)
• The maximum data value up to that point in the report (SUBMAX)
• The maximum data value in the section (MAX)
• The average data value up to that point in the report (SUBAVG)
• The average data value in the section (AVG)

Each value will be preceded, on the same line, by appropriate identifying text. Two
columns of data will be printed.
To print the report, the following is coded:

SORT FIELDS=(1,3,CH,A,5,2,CH,A) SORT DATA BY GROUP AND SECTION

OUTFIL FILES=(OUT),
SECTIONS=(1,3,SKIP=3L,
HEADER3=(3:'GROUP',2X,1,3,/,16:'SECTION',6X,'VALUE',/),
TRAILER3=(//,4:'MINIMUM VALUE TO THIS POINT= ',
35:SUBMIN=(8,6,ZD,M2),/,
4:'MINIMUM VALUE FOR THIS GROUP= ',
35:MIN=(8,6,ZD,M2),//,
4:'MAXIMUM VALUE TO THIS POINT= ',
35:SUBMAX=(8,6,ZD,M2),/,
4:'MAXIMUM VALUE FOR THIS GROUP= ',
35:MAX=(8,6,ZD,M2),//,
4:'AVERAGE VALUE TO THIS POINT= ',
35:SUBAVG=(8,6,ZD,M2)/,
4:'AVERAGE VALUE FOR THIS GROUP= ',
35:AVG=(8,6,ZD,M2))),
OUTREC=(18:5,2,26:8,6,ZD,M2,80:1X)

Figure 236. Sample Code to Print Report

The following shows two sections from the report, with the resulting values for
subminimums, minimums, submaximums, maximums, subaverages and averages:

Syncsort MFX Programmer’s Guide 3–57

 GROUP AAA
SECTION VALUE

01 38.42
01 923.12
01 8,756.33
02 9,723.63
02 67.43
02 175.66
03 645.83
03 673.41
03 23.71
MINIMUM VALUE TO THIS POINT= 23.71
MINIMUM VALUE FOR THIS GROUP= 23.71

MAXIMUM VALUE TO THIS POINT= 9,723.63

MAXIMUM VALUE FOR THIS GROUP= 9,723.63

AVERAGE VALUE TO THIS POINT= 2,336.39

AVERAGE VALUE FOR THIS GROUP= 2,336.39
GROUP BBB
SECTION VALUE

01 0.01
01 456.11
01 874.01
02 4,354.00
02 2,583.54
02 3.57
03 809.01
03 934.53
03 853.21
MINIMUM VALUE TO THIS POINT= 0.01
MINIMUM VALUE FOR THIS GROUP= 0.01

MAXIMUM VALUE TO THIS POINT= 9,723.63

MAXIMUM VALUE FOR THIS GROUP= 4,354.00

AVERAGE VALUE TO THIS POINT= 1,771.97

AVERAGE VALUE FOR THIS GROUP= 1,207.55

Figure 237. Sample Report Sections

Explanation: The SECTION parameter generates a section break on field 1,3,

which identifies data groups (AAA, BBB, etc.). The HEADER3 parameter defines
section headers that print the label "GROUP" followed by the data group identifier.
HEADER3 also defines two column headings: "SECTION," which identifies the
column containing section numbers, and "VALUE," which identifies the columns
containing the numeric data.

3–58 Syncsort MFX Programmer’s Guide

 The TRAILER3 subparameters are SUBMIN, MIN, SUBMAX, MAX, SUBAVG and
AVG. They specify the six values to appear in the section trailer. The values are all
derived from the same field (8,6) and are suitably edited with mask M2
(8,6,ZD,M2).
The OUTREC parameter places the two data fields (5,2 and 8,6) in the report and
edits the 8,6 field in the same way as for the six values in the section trailer
(8,6,ZD,M2). The blank space placed at position 80 (80:1X) ensures that the output
record is long enough to contain the header records.

Counting Data Records

Trailers in a report will sometimes require you to obtain a record count or a count
for a particular type of item in a specific part of a report. The OUTFIL statement
allows you to write trailers that contain such a count as well as cumulative, or
running, counts of records. Moreover, you can obtain these counts at the end of a
report, at the end of a page, and at the end of a section.
To generate these counts, use the COUNT and SUBCOUNT subparameters (or
COUNT15 and SUBCOUNT15). These subparameters can be used in conjunction
with all other TRAILER entries. For syntax of COUNT and SUBCOUNT (as well
as COUNT15 and SUBCOUNT15), see “TRAILER Parameters (Optional)” on
page 2-105.

Obtaining a Count of Data Records

Example: Marketing wants a count of the total number of customers with
outstanding payments included in the summary of its outstanding invoices report.
To get this record count and print it as part of the report summary, the following is
coded.

Syncsort MFX Programmer’s Guide 3–59

 //INVLST JOB Gives the Jobname
// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=INVOICE,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,23,CH,A) Sorts Records
.
.
.
OUTFIL.
.
.
TRAILER1=(20/, Generates Report Summary
40:'NUMBER OF CUSTOMERS WITH OUTSTANDING PAYMENTS:',
COUNT)

Figure 238. JCL and Required Control Statements

Figure 239 shows the trailer containing the record count.

NUMBER OF CUSTOMERS WITH OUTSTANDING PAYMENTS: 52

Figure 239. Report Trailer Containing Record Count

Explanation: Since each record in the report represents an individual customer,

coding the COUNT entry in the TRAILER1 will provide the total number of
customers with outstanding payments. This TRAILER1 produces a report trailer,
or summary, that constitutes the final page of a report. It will print on the 21st line
of the page (20/) and begin printing the literal string 'NUMBER OF CUSTOMERS
WITH OUTSTANDING PAYMENTS: ' in column 40.

Obtaining a Cumulative (Running) Count of Data Records

Example: For an outstanding invoices report sectioned by month, marketing wants
a cumulative, or running, count of invoices to date at the end of each section as well
as a total count of each month’s invoices included as section trailers.

3–60 Syncsort MFX Programmer’s Guide

 To generate these record counts, the following is coded.

//INVLST JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=INVOICE,DISP=SHR Defines Input Data Set
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD SPACE=(CYL,5),UNIT=SYSDA Defines Intermediate Storage
//SYSIN *
SORT FIELDS=(28,2,ZD,A, Sorts Records
24,2,ZD,A,
1,23,ZD,A)
.
.
.
OUTFIL.
.
.
SECTIONS=(24,6,SKIP=1L, Generates Sections with Record
TRAILER3=(/
, Count & Cumulative Record Subcount
95:'MONTH''S NUMBER OF INVOICES: ',COUNT,/,
95:'NUMBER OF INVOICES TO DATE: ',SUBCOUNT))

Figure 240. JCL and Required Control Statements

Syncsort MFX Programmer’s Guide 3–61

 Figure 241 shows the trailers containing the counts of records.

. . . .
. . . .
. . . .
RIBBIT TECHNOLOGIES 2/15/91 360.00 21.60
RICE FEATURES 12/15/91 750.00 75.00
SIDNEY COLLEGE 12/15/91 5,000.00 300.00
SNAP FEATURES 12/15/91 750.00 75.00
WEBB BROS CORP 12/15/91 600.00 36.00
WELLINGTON IMPORTS 12/15/91 750.00 45.00
WINIFRED INDUST 12/15/91 350.00 26.00

MONTH'S NUMBER OF INVOICES: 17

NUMBER OF INVOICES TO DATE: 17

ARLINE FRAGRANCES 1/17/92 7,500.00 618.75

CHARACTER DATA 1/17/92 1,100.00 50.75
COUNTRY INDUSTRIAL 1/17/92 850.00 0.00
DUNHAM INDUST CO 1/17/92 850.00 0.00
ECHO LABS INC 1/17/92 550.00 22.00
ESS SECURITIES 1/17/92 550.00 22.00
EVERMORE INDUST 1/17/92 3,000.00 225.00
GOODEY FOODS 1/17/92 600.00 30.00
GROSS BOOKS CO 1/17/92 2,500.00 150.00
HARVEY MOTORS CO 1/17/92 3,000.00 225.00
KALABRA CORP 1/17/92 1,500.00 90.00
MEWER COLLEGE 1/17/92 1,500.00 75.00
NORTHEAST INDUST 1/17/92 200.00 20.00
REPUBLIC DATA 1/17/92 1,100.00 90.75
RICE FEATURES 1/17/92 1,500.00 75.00
ROBINS NEST CORP 1/17/92 900.00 54.00
SIDNEY COLLEGE 1/17/92 2,500.00 150.00
SOUTHWEST INDUST 1/17/92 200.00 20.00
SPENSERS INDUST 1/17/92 650.00 26.00
UNITED INTERESTS 1/17/92 1,500.00 90.00
WINIFRED INDUST 1/17/92 650.00 26.00

MONTH'S NUMBER OF INVOICES: 21

NUMBER OF INVOICES TO DATE: 38

BALTIC AVENUE CORP 2/15/92 650.00 29.25

BATHO PRODUCTS 2/15/92 850.00 51.00
CARRINGTON OIL 2/15/92 1,600.00 64.00
CDR TRUST INC 2/15/92 1,500.00 75.00
ECHO LABS INC 2/15/92 550.00 22.00
ESS SECURITIES 2/15/92 550.00 22.00
FASTEROOT EQUIP 2/15/92 1,700.00 76.50
FEDERAL FABRICS 2/15/92 1,750.00 70.00
. . . .
. . . .
. . . .

Figure 241. TRAILER3 Containing Record Counts and Cumulative Record Counts

Explanation: The trailer’s first / entry causes the printer to leave one blank line
after the data records and before printing the trailer. The second / entry indicates
the end of the trailer’s first line. The identical number-colon entries (95:) set the
starting positions of the literal strings that follow them: 'MONTH' 'S NUMBER OF

3–62 Syncsort MFX Programmer’s Guide

 INVOICES: ' and 'NUMBER OF INVOICES TO DATE: '.(Note that the apostrophe
in MONTH'S is doubled because a single apostrophe would signal the end of a
literal string.) Finally, because each data record in this report represents an
invoice, the TRAILER3’s COUNT entry generates a count of each month’s invoices
and the SUBCOUNT entry generates a cumulative, or running, count of the
invoices. The leading zeros in these 8-byte fields are suppressed.

Creating Multiple Output Files

Data centers often use the same masterfile for different purposes. Assume, for
example, that you wanted to produce two reports using a masterfile of cash-receipt
records. One report was to present the total cash receipts for the current month;
the second, for the year to date. This would typically entail running a separate sort
for each report. SortWriter’s multiple-output feature, however, enables you to
produce both reports with a single pass of the sort. In addition, you can specify the
same or different devices to receive the separate output files.
Note: All the output files will be sequenced in the same way, as specified on the
SORT or MERGE statement. If you need to sort the output files differently, you
should use Syncsort PipeSort, a Precisely product that works with Syncsort MFX to
reduce total elapsed time by generating multiple, differently sequenced output files
from a single read of the input data.
To generate multiple output files, code the OUTFIL statement. For syntax of the
OUTFIL control statement, see “OUTFIL Control Statement” on page 2-85.

Generating Several Output Files with Different Information

Example: Marketing wants three output files of customer records. The first will
contain a list of U.S. and European customers. The second will contain a list of U.S.
customers only, and the third will contain a list of European customers only.

Syncsort MFX Programmer’s Guide 3–63

 To generate the three separate files, the following is coded.

//CUSTRCD JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=A Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=SALES.RECORDS, Defines Input Data Set
// VOL=SER=DISK1,
// DISP=SHR
//SORTOF1 DD DSN=SORTED.CUSTM.RECORDS, Defines First Output Data
// UNIT=TAPE,VOL=SER=112231, Set Containing All
//* Customer Records
// DISP=(NEW,KEEP)
//SORTOF2 DD DSN=SORTED.DCUSTM.RECORDS, Defines Second Output Data
Set Containing Domestic
// UNIT=TAPE,VOL=SER=112232, Customer Records Only
// DISP=(NEW,KEEP)
//SORTOF3 DD DSN=SORTED.ECUSTM.RECORDS, Defines Third Output Data
//* Set Containing European
// UNIT=TAPE,VOL=SER=112233, Customers Only
// DISP=(NEW,KEEP)
//SORTWK01 DD SPACE=(CYL,20),UNIT=SYSDA Defines Intermediate Storage
//SORTWK02 DD SPACE=(CYL,20),UNIT=SYSDA Defines Intermediate Storage
//SORTWK03 DD SPACE=(CYL,20),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(10,15,CH,A) Sorts Records
OUTFIL FILES=1, OUTFIL Statement for SORTOF1
INCLUDE=ALL Including All Records
OUTFIL FILES=2, OUTFIL Statement for SORTOF2
INCLUDE=(67,3,CH,EQ,C'USA') Including USA Records
OUTFIL FILES=3, OUTFIL Statement for SORTOF3
INCLUDE=(67,3,CH,EQ,C'EUR') Including Eur. Records

Figure 242. JCL and Required Control Statements

Explanation: Creating the three requested output files requires coding three
SORTOFxDD statements in the JCL: SORTOF1, SORTOF2, and SORTOF3 as well
as three OUTFIL statements. Each of the OUTFIL statements is connected by a
FILES parameter to one of the output files defined in the JCL. Specifying 1 on the
FILES parameter connects its OUTFIL statement with the output file defined by
the SORTOF1 DD statement in the JCL. Likewise, specifying 2 connects its
OUTFIL statement with the output file defined by SORTOF2, and so on. The first
output file will contain all the records from the input file (INCLUDE=ALL). The
second output file will include only those records that contain the character string
'USA' beginning in byte 67, (INCLUDE=(67,3,CH,EQ,C'USA')), which indicates
that these records are for USA customers. And similarly, the third output file will
include only those records that contain the character string 'EUR' beginning in
byte 67, which indicates that these records are for European customers.

3–64 Syncsort MFX Programmer’s Guide

Writing Identical Output Files to Different Devices
Example: Personnel wants a printed copy of its updated masterfile as well as copies
on disk and on tape.
To generate these three copies of the same file on different devices, the following is
coded.

//MULTOUT JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD DSN=PERSNL.RECORDS, Defines Input Data Set
// VOL=SER=DISK1,
// DISP=SHR
//SORTOFPR DD SYSOUT=* Defines Printed Output
//* Data Set
//SORTOFTP DD DSN=PERSNL.RECORDS.TAPE, Defines Tape Output Data Set
// UNIT=TAPE,VOL=SER=112233,
// DISP=(NEW,KEEP)
//SORTOFDS DD DSN=PERSNL.RECORDS.DISK, Defines Disk Output Data Set
// UNIT=DISK1,DISP=(NEW,KEEP),
// SPACE=(CYL,60)
//SORTWK01 DD SPACE=(CYL,20),UNIT=SYSDA Defines Intermediate Storage
//SORTWK02 DD SPACE=(CYL,20),UNIT=SYSDA Defines Intermediate Storage
//SORTWK03 DD SPACE=(CYL,20),UNIT=SYSDA Defines Intermediate Storage
//SYSIN DD *
SORT FIELDS=(1,40,CH,A) Sorts Records
OUTFIL FILES=(PR,TP,DS) Creates Multiple Output

Figure 243. JCL and Required Control Statements

Explanation: Creating the three copies of the updated masterfile requires coding
only one OUTFIL statement with a FILES parameter. The FILES parameter
instructs MFX to look for multiple output files defined in the JCL and to send its
output to the devices specified in the SORTOFxx statements. Thus, the output that
has been sorted as specified on the SORT statement (1,40,CH,A) will be sent to the
printer specified in the SORTOFPR statement, to the tape volume specified in the
SORTOFTP statement, and to the disk data set specified in the SORTOFDS
statement.

Syncsort MFX Programmer’s Guide 3–65

E-mailing a Report in PDF Format
MFX’s output can be created as a PDF file and optionally sent as an e-mail
attachment. In this example, there are two OUTFIL statements that each create a
simple one-page report.
The OUTREC parameter of each OUTFIL reformats the input records to add
spacing between fields and reformats a ZD field as a dollar amount. HEADER2 and
TRAILER2 parameters add a page header and trailer with a date and total
amount.
The first OUTFIL statement writes this report to the default SORTOUT DD
statement SYSOUT output. The second OUTFIL statement writes the same report,
but adds the OUTPUT parameter and EMAIL subparameter, directing it to the
SORTOF1 DD statement, which defines an HFS file. The OUTPUT parameter
requests creating the SORTOF1 file as a PDF file. (This is the default for OUTPUT.
Optionally, an HTML or RTF file could have been created.) OUTPUT also
establishes a 144-point (2”) left margin, a background color of LIGHTGRAY, and
different fonts and colors for the header (FONTH2), trailer (FONTT2), and detail
lines (FONT). The EMAIL subparameter requests that an e-mail be sent to a
named recipient (TO) and to a list of recipients defined in the EMAILDD DD
statement (TODD), with the file sent as an e-mail attachment. FROM and
SUBJECT are also defined for the e-mails.
The LINES parameter sets 38 lines per page, which is the proper number of report
lines that will fit on a PDF page of the default LETTER page size when using a 12-
point font and the default 36-point top and bottom margins.

3–66 Syncsort MFX Programmer’s Guide

 On the following pages, Figure 244 contains the JCL and control statements
required to produce the generated PDF report in Figure 245.

//EMAILPDF JOB Gives the Jobname

// EXEC PGM=SYNCSORT Identifies the Program
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//SORTIN DD * Input
Joseph Smith 02506240
James Jones 12345678
John Jackson 00987654
Mary Lee 07677201
Michael Jay 04216797
//SORTOUT DD SYSOUT=* Basic printed report
//SORTOF1 DD PATH=’/u/sales_report.pdf’,PATHMODE=SIRWXU,
// PATHOPTS=(ORDWR,OCREAT) Enhanced PDF report
//EMAILDD DD * Email address list, end with “;”
jack@anycompany.com;
jill@anycompany.com;
//STDERR DD SYSOUT=* Java error messages
//STDOUT DD SYSOUT=* Java messages
//SYSIN DD *
SORT FIELDS=(21,7,ZD,D) sort from highest to lowest sales
OUTFIL OUTREC=(1,20,24:C’$’,21,8,ZD,M2,LENGTH=13), printed report
HEADER2=(‘September Sales Report ‘,&DATE,3/,
‘Salesperson Name’,24:’ Sales’,/),
TRAILER2=(‘Total Sales:’,24:’$’,TOT=(21,8,ZD,M2,LENGTH=13))
OUTFIL OUTREC=(1,20,24:C’$’,21,8,ZD,M2,LENGTH=13), PDF report
HEADER2=(‘September Sales Report ‘,&DATE,3/,
‘Salesperson Name’,24:’ Sales’,/),
TRAILER2=(‘Total Sales:’,24:’$’,TOT=(21,8,ZD,M2,LENGTH=13)),
OUTPUT=(MARGINS=(LEFT=144PT), 2 inch left margin
BACKGROUNDCOLOR=LIGHTGRAY, pleasing background
FONTH2=(BOLD,RED), highlight page titles
FONT=BLUE, color for detail lines
FONTT2=(BLACK,UNDERLINE), underline summary line
EMAIL=(TO=’sales_manager@anycompany.com’, mgr email
TODD=EMAILDD, staff emails
FROM=’IT_dept@anycompany.com’, sender email
SUBJECT=’New Sales Report’)), email subject
LINES=38, 12pt PDF font ==> 38 line/pg
FILES=1 SORTOF1 DD defines HFS file
/*

Figure 244. JCL and Required Control Statements

Syncsort MFX Programmer’s Guide 3–67

 Figure 245. PDF Report

3–68 Syncsort MFX Programmer’s Guide

Chapter 4 JCL and Sample JCL/
Control Statement Streams

MFX’s job control statements follow the standard operating system conventions
described in the z/OS job control language manuals. Each program application
therefore requires a JOB statement, an EXEC statement, and a DD (data
definition) statement for every data set used. (The single exception to this is the
dynamic allocation of work files via DYNALLOC or DYNATAPE.) The inclusion
and coding requirements of particular job control statements depend on such
factors as whether MFX is program-invoked or initiated directly, whether any exits
are coded, and, of course, whether the sorting technique requested is Disk Sort,
MAXSORT, or PARASORT.
All aspects of program initiation which are specific to the sort/merge (such as the
dedicated DD names SORTIN and SORTOUT) are documented in this chapter. For
complete coding instructions, refer to a z/OS MVS JCL reference manual.
The following table summarizes MFX’s DD statement requirements.

Syncsort MFX Programmer’s Guide 4–1

 //STEPLIB DD Instructs operating system to look for the sort program in a
//JOBLIB DD specified data set.
//SYSOUT DD Message data set. Required unless all messages are routed to
console.
//SORTIN DD Defines the input data set for a SORT or COPY application.
Required unless it is a join application or there is an E15 exit
routine. Ignored if the invoking program supplies an inline
E15 exit routine; optional if the MODS statement activates
an E15 exit routine.
//SORTINnn DD MERGE input data set. Required unless there is an E32.
//SORTINn DD SORTINnn DD statements are not processed when
FIELDS=COPY is specified.
//SORTJNF1 DD Join application input data sets. Required for join
//SORTJNF2 DD application.
//SORTOUT DD Output data set. Required unless there is an E35. Ignored if
the invoking program supplies an inline E35 exit routine;
optional if the MODS statement activates an E35 exit rou-
tine.
//SORTOFxx DD OUTFIL output data sets. One required for each FILES or
//SORTOFx DD FNAMES specification.
//fname DD
//SORTXDUP DD Output data set of records eliminated by the DUPKEYS
control statement. Required when the XDUP parameter is
specified.
//SORTXSUM DD Output data set of records eliminated by the SUM control
statement. Required when the XSUM parameter is specified.
//SORTWKxx DD Disk work area definition. Required unless incore sort,
//SORTWKn DD DYNALLOC, MERGE, COPY or restarting at a MAXSORT
merge breakpoint.
//SYSIN DD Control statement data set. Required unless control state-
ments are supplied via an invoking program parameter list.
//$ORTPARM DD Used to override PARM or control statement information.
//SORTCKPT DD Checkpoint data set. Required for Checkpoint-Restart.
//SORTMODS DD Required if user exits are in SYSIN.
//SYSLIN DD Required if user exits are to be linkage-edited at execution
//SYSLMOD DD time.
//SYSPRINT DD
//ddname DD Required for exits unless the exit is inline in LINKLIB/JOB-
LIB/STEPLIB or in SYSIN.
//STDOUT DD Data sets for informational and error messages issued by
//STDERR DD Java during creation of PDF, HTML and RTF files via the
OUTFIL OUTPUT facility.

Table 44. MFX DD Statements

4–2 Syncsort MFX Programmer’s Guide

EXEC Statement
The EXEC statement is required in order to indicate to the operating system that
the job is a sort/merge application. For a Disk Sort, the format of the EXEC
statement is as follows.

PGM=SYNCSORT
 
 PGM=SORT 
 
//stepname EXEC  PGM=IERRCO00  [,PARM='...']
 
 PGM=IGHRCO00 
 PGM=ICEMAN 

Figure 246. Disk Sort EXEC Statement Format

To use a sort cataloged procedure, omit PGM= and specify the appropriate
procedure name.
The PARM parameter may be used to pass the sort/merge program a variety of
keyword parameters, modifying it to meet the needs of the individual application.

For MAXSORT, PARASORT, DB2 Query and MULTIIN Sup-

port
The format of the EXEC statement varies with the sorting technique chosen. The
MAXSORT and PARASORT PARM options are used to request the MAXSORT or
PARASORT sorting technique. The DB2 PARM option is used to request the DB2
Query function. The MULTIIN PARM is used to request the MULTIIN facility.

Coding Conventions for DD Statements

The following table summarizes the standard coding conventions for DD
statements as they relate to the sort/merge program. For more detailed
information, refer to a z/OS Job Control Language manual.

Syncsort MFX Programmer’s Guide 4–3

 Parameter Subparameter Required?
DSNAME/DSN To access a labeled data set (e.g., SORTIN,
STEPLIB) or to keep or catalog the data set
being created (e.g., SORTOUT, SORTOU00).
DCB RECFM, LRECL, DCB not required for disk or standard labeled
and BLKSIZE tape input.

OPTCD and To override the values in the data set label of

BUFOFF an old data set; to override the values in the
first SORTIN or SORTINnn file for a new data
set.

To indicate ASCII input and output.

UNIT For an input file that is not cataloged or
passed; for a new data set
SPACE For a new DASD data set.
VOLUME/VOL For an input file that is not cataloged or
passed; for a DASD output data set to be cata-
loged or passed.
LABEL To override (1,SL).
DISP To override (NEW,DELETE).

Table 45. DD Statement Parameters, Standard Coding Conventions

STEPLIB/JOBLIB DD Statement
If MFX has been installed in a private user library or in a test library, a STEPLIB
or JOBLIB DD statement is required. The sample DD statement below instructs
the operating system to look for the sort in a partitioned data set named
SYNCTEST.

//STEPLIB DD DSN=SYNCTEST,DISP=SHR

Figure 247. Sample STEPLIB DD Statement

4–4 Syncsort MFX Programmer’s Guide

SYSOUT DD Statement
This defines the data set for MFX messages.

//SYSOUT DD SYSOUT=A

Figure 248. Sample SYSOUT DD Statement

If the SYSOUT DD statement is omitted, any message routed to it will be diverted

to the console. Omitting the SYSOUT DD statement and setting the MSG=SC
PARM (critical messages to the console, all messages to the printer), for example,
will result in all messages being sent to the console.

SORTIN DD Statement
The SORTIN DD statement defines the data set(s) to be sorted or copied. (The
input files for a merge application are defined by the SORTINnn DD statement.) It
is required for all sorts except those where an E15 exit (COBOL Input Procedure)
provides all the input records or where the MULTIIN facility or join facility is used.
The MULTIIN facility is used to combine VSAM and non-VSAM data sets as input
to a SORT or COPY, and SORTMInn DD statements are used in place of SORTIN.
The join feature joins records from two input files that are specified on the
SORTJNF1 and SORTJNF2 DD statements.
The SORTIN file must have physical sequential or extended sequential
organization or be a member of a partitioned data set or PDSE. It may reside on
any device supported by BSAM or VSAM and if it is a VSAM data set, may be key-
sequenced, entry-sequenced or relative record. SORTIN data sets may also be
BatchPipes or z/OS pipes or they may be HFS data sets. DCB information need not
be supplied for a disk or standard labeled tape file. Any of the information accessed
from a standard label can be overridden by coding the appropriate DCB parameter
in the JCL.
The maximum record lengths supported are 32,760 bytes for fixed-length records
and 32,767 bytes for variable-length records.
By default MFX does not accept an uninitialized SORTIN data set and will
terminate processing with a WER400A message. An uninitialized data set is one
that has been newly created but never successfully closed. The UNINTDS PARM or
installation option can be used to change MFX’s default mode of processing to
accept an uninitialized input data set and process it as an empty file. See
“UNINTDS” on page 5-30.

Syncsort MFX Programmer’s Guide 4–5

 In this example, the data set to be sorted/copied is named SALESIN. It resides on
one reel of tape whose volume serial number is 123456. SALESIN is the first data
set on that tape and has a standard label.

//SORTIN DD DSN=SALESIN,DISP=(OLD,KEEP),
// UNIT=TAPE,VOL=SER=123456

Figure 249. Sample SORTIN DD Statement

To access a SORTIN data set that resides in hiperbatch use the HBSI PARM. For
more information about HBSI see “Chapter 5 PARM Options”.
Note: The TYPE parameter on the RECORD control statement should be specified
if SORTIN is VSAM. If TYPE is not provided, the SORTOUT RECFM will be
examined to determine the SORTIN TYPE. If no SORTOUT RECFM is found,
TYPE=V will be assumed if the SORTOUT is VSAM and TYPE=F if the SORTOUT
is non-VSAM.
Note: RLS mode (RLS=CR and RLS=NRI) or Linear VSAM data sets are not
supported for input or output.

Concatenating Input Data Sets

The SORTIN file may consist of concatenated data sets, up to the limit supported
by the operating system. (Note: If you want to combine VSAM and non-VSAM files
as input to a SORT or COPY, use the MULTIIN facility described in Chapter 12.)
MFX must determine one set of DCB characteristics to use for reading all data sets
in the concatenation. The following rules apply to the DCB characteristics:
• When the first data set is fixed-length (RECFM=F, FB, FBS), all subsequent
data sets must be fixed-length and have the same LRECL.
• When the first data set is variable-length (RECFM=V, VB, VS, VBS), all
subsequent data sets must be variable-length.
• For variable-length data sets, the LRECL of the first data set is used except for
the following situations:
• The LRECL of a subsequent data set is used if that LRECL is the largest
found and is available at sort initialization. An LRECL is available at
initialization if it is specified on a SORTIN DD statement or exists in the
label of a SORTIN disk data set.
• A record length specified via the l1 value on the RECORD control statement
is used if it is the largest record length found.
• For both fixed and variable-length data sets, the BLKSIZE of the first data set
is used unless the BLKSIZE of a subsequent data set is the largest found and is
available at sort initialization. A BLKSIZE is available at initialization if it is
specified on a SORTIN DD statement or exists in the label of a SORTIN disk
data set.

4–6 Syncsort MFX Programmer’s Guide

 The following shows sample JCL for concatenating input data sets:

//SORTIN DD DSN=AUGUST.SALES,DISP=(OLD,KEEP),
// UNIT=3390,VOL=SER=DISK1,
// DCB=(LRECL=200,RECFM=VB,BLKSIZE=7404)
// DD DSN=JUNE.SALES,DISP=(OLD,KEEP),
// UNIT=TAPE,VOL=SER=123456,LABEL=(2,SL),
// DCB=(LRECL=200,RECFM=V,BLKSIZE=8004)
// DD DSN=JULY.SALES,DISP=(OLD,KEEP),
// UNIT=TAPE,VOL=SER=654321,LABEL=(1,SL),
// DCB=(LRECL=100,RECFM=VB,BLKSIZE=8004)

Figure 250. Sample Disk and Tape Data Set Concatenation to SORTIN

In the preceding example, one disk and two tape data sets have been concatenated.
Any one of these data sets could be presented first. Position is not dependent upon
BLKSIZE or LRECL. If the LRECL or BLKSIZE cannot be determined at SORT
initialization, the first data set must carry the largest LRECL or BLKSIZE of the
concatenation. Typically the LRECL or BLKSIZE cannot be determined when the
input consists of concatenated tape data sets and the JCL lacks a DCB
specification.

Sorting Large Input Data Sets

The MAXSORT technique is recommended for sorting very large amounts of data
when disk work space is limited. With this technique, SORTWK requirements are
independent of SORTIN size; thus, regardless of the size of the file, it can be sorted
by one sort program using disk work files. MAXSORT’s breakpoint/restart
capability breaks the overlarge sorting application into smaller individual sorts;
high priority jobs can execute between these smaller sorts without forcing any data
to be resorted. See “Chapter 9 MAXSORT”.

Reducing Elapsed Time for SORTS with Multi-volume or Concate-

nated Tape SORTIN
The PARASORT technique can be used to improve elapsed time performance of
sorts that use multi-volume or concatenated tape SORTIN data sets. (See “Chapter
10 PARASORT”.)

SORTINnn or SORTINn DD Statement

SORTINnn and SORTINn DD statements are used to define the input to a merge
application. (Use the SORTIN DD statement to define the data set to be sorted or
copied.) SORTINnn or SORTINn DD statements are required for all merge
applications unless an E32 exit supplies the input data. SORTINnn and SORTINn

Syncsort MFX Programmer’s Guide 4–7

 data sets may be BatchPipes or z/OS pipes or they may be HFS data sets. Since all
input data sets are open at the same time during a merge, UNIT=AFF cannot be
coded on any of the input DD statements.
It is possible to merge up to 100 data sets. Each input data set is specified on a
SORTINnn or SORTINn DD statement. The valid range for n is 0 through 9; for
nn, 00 through 99. If both SORTINx and a SORTIN0x are specified, they are
treated as duplicates and only the first definition is processed. Each file must
receive a different number. Numbers may be skipped or used out of order. There are
no restrictions as to which input files are to receive which numbers.
All input data sets must have the same record format (fixed or variable), and the
records in each input file must already be in the desired sequence.
By default, MFX does not accept an uninitialized SORTINnn or SORTINn data set
and will terminate processing with a WER400A message. An uninitialized data set
is one that has been newly created, but never successfully closed. The UNINTDS
PARM or installation option can be used to change MFX’s default mode of
processing to accept an uninitialized input data set and process it as an empty file.
(See “UNINTDS” on page 5-30.)

//SORTIN17 DD DSNAME=BRANCHA.FICA,VOL=SER=131313,
// DISP=OLD,UNIT=3480
//SORTIN01 DD DSNAME=BRANCHC.FICA,VOL=SER=242424,
// DISP=OLD,UNIT=3390
//SORTIN24 DD DSNAME=BRANCHB.FICA,VOL=SER=121212,
// DISP=OLD,UNIT=3400-3,LABEL=(,NL),
// DCB=(RECFM=FB,LRECL=80,BLKSIZE=400)

Figure 251. Sample SORTINnn DD Statements (Merge)

In this example, the DCB information for the first two of the three files to be
merged is supplied by the file labels. In order for the merge to execute, these files
must have a RECFM of F or FB, as indicated by the third file’s RECFM value.

4–8 Syncsort MFX Programmer’s Guide

SORTJNF1 and SORTJNF2 DD Statements
SORTJNF1 and SORTJNF2 DD statements are used to define the input to a join
application. These statements are required for a join application.
In this example, the SORTJNF1 and SORTJNF2 files are specified. The first file is
a fixed-length file; the second a variable-length file.

//SORTJNF1 DD DSNAME=BRANCHA.FICA,VOL=SER=131313,
// DISP=OLD,UNIT=3480,
// DCB=(RECFM=FB,LRECL=80,BLKSIZE=8000)
//SORTJNF2 DD DSNAME=MASTER.FICA,VOL=SER=121212,
// DISP=OLD,UNIT=3390,
// DCB=(RECFM=VB,LRECL=200,BLKSIZE=27800)

Figure 252. Sample SORTJNF1 and SORTJNF2 DD Statements

SORTOUT, SORTOFxx, SORTOFx, SORTXSUM, and

SORTXDUP DD Statements
The SORTOUT, SORTOFxx, SORTOFx, SORTXSUM, and SORTXDUP DD
statements are used to define one or more output files. The FNAMES parameter of
the OUTFIL control statement may also specify DD names of output files. All
output is directed to SORTOUT unless an inline E35 exit (COBOL output
procedure) assumes the full responsibility for output processing. Records
eliminated by SUM processing will be written to the SORTXSUM DD statement if
the XSUM option was selected on the SUM control statement. Records eliminated
by DUPKEYS processing will be written to the SORTXDUP DD statement if the
XDUP option was selected on the DUPKEYS control statement. These output data
sets may be directed to a BSAM or VSAM supported device, to BatchPipes or z/OS
pipes, or to HFS data sets.

//SORTOUT DD DSN=MASTER.OUT,UNIT=SYSDA,
// DISP=(NEW,KEEP),SPACE=(TRK,10),
// VOL=SER=DSK002
//SORTOF01 DD DSN=REPORT.OUT,UNIT=SYSDA,
// DISP=(NEW,KEEP),SPACE=(TRK,10),
// VOL=SER=DSK002

Figure 253. Sample SORTOUT/SORTOFxx DD Statements

In the preceding example, the missing DCB parameters except BLKSIZE will
default to those assigned to SORTIN or (for a merge application) to those assigned
to the last SORTINnn in the JCL stream. The DCB BLKSIZE, if missing, will be

Syncsort MFX Programmer’s Guide 4–9

 determined via system-determined blocksize when it is active or from SORTIN if
SORTOUT and SORTIN LRECLs are the same, otherwise MFX will select an
appropriate BLKSIZE.
If a sort or a merge has an LRECL specified in the output DD JCL that is found to
be smaller than the internally processed record length (determined from SORTIN,
the LENGTH values of a RECORD statement, or an INREC statement), MFX
processing will be controlled by the SOTRN installation option or its run-time
override parameter TRUNC. (SYNCGENR applications are controlled by the
SOTRNGN installation option.) If the parameter setting allows truncation, MFX
will write the records to the output data set by truncating the records to the
LRECL of that data set. The delivered default allows truncation. MFX will not
truncate records after OUTREC processing. If the option disallows truncation, a
WER462A error message will be issued.
If an application that is processing fixed-length data has an LRECL specified in the
SORTOUT or SORTXSUM JCL that is found to be longer than the internally
processed record length, MFX will normally pad the output records with binary
zeros. See the discussion of the PAD parameter in chapter 5 for additional controls
that can be applied to applications with both a SORTIN and a SORTOUT where
the SORTOUT LRECL is longer than the SORTIN LRECL. This padding will be
done for SORTXSUM and for SORTOUT when OUTFIL is not in use. It will not be
done for any OUTFIL files. If the option disallows padding, a WER462A error
message will be issued. The delivered default allows padding.
If RECFM is specified and the report writing features of the OUTFIL control
statement are being used, the RECFM of the output file must include the 'A'
subparameter, except when the REMOVECC parameter is in use.
For a COPY or MERGE, the output file must not be the same as any of the input
files.
Note: The TYPE parameter on the RECORD control statement should be specified
if SORTIN is VSAM. If TYPE is not provided, the SORTOUT RECFM will be
examined to determine the SORTIN TYPE. If no SORTOUT RECFM is found,
TYPE=V will be assumed if the SORTOUT is VSAM and TYPE=F if the SORTOUT
is non-VSAM.
Note: RLS mode (RLS=CR and RLS=NRI) or Linear VSAM data sets are not
supported for input or output.

Secondary Allocation
If the automatic secondary allocation option was enabled at installation time,
requesting secondary allocation on the output DD statements is not required. This
feature automatically provides output space for each of the output files.
To place a SORTOUT data set into hiperbatch so that subsequent job steps can
access it, use HBSO. For more information about HBSO see the PARM Option
chapter in this manual.

4–10 Syncsort MFX Programmer’s Guide

SORTWKxx or SORTWKx DD Statement
For non-MAXSORT applications, up to 255 data sets may be specified for
intermediate storage when sorting. (MAXSORT, which is recommended for large
sorting applications, is limited to 32 SORTWK data sets.) Each work file carries a
SORTWKxx or SORTWKx name.
x can be any alphanumeric or national ($, #, @) character. Each SORTWKxx or
SORTWKx must be allocated on a single unit and a single volume and should have
a unique name. For example, SORTWK01, SORTWK02, etc. SORTWK data sets
must have physical sequential organization and cannot be extended sequential. For
performance reasons, the use of VIO for SORTWK data sets is not recommended.
You can specify 3380 and/or 3390 disk devices. When device types are mixed, each
device is used to full capacity. Note that although SORTWK space can be allocated
in blocks, tracks, or cylinders, allocating in cylinders will yield optimal
performance. The CONTIG option of the SPACE parameter should be avoided since
it may delay allocation and offers no performance advantage.
The SORTWKxx DD statement in the following example establishes a primary
allocation of 20 cylinders of work space.

//SORTWK02 DD UNIT=3390,SPACE=(CYL,20)

Figure 254. Sample SORTWKxx DD Statement for Disk Sorts

Secondary Allocation
There is no need to specify RLSE and a secondary allocation value on the
SORTWKxx DD statement at installations that have set these defaults at MFX
installation time.

Are SORTWKxx DD Statements Necessary?

SORTWKxx DD statements are not used for merge or copy applications. They are
not required for sorts executed using the DYNALLOC option. Provided neither
DYNALLOC nor FIELDS=COPY is in effect, it will be necessary to include
SORTWK data sets whenever any of these conditions holds:
• INCORE is set to OFF.
• An E14 or E16 is included.
• Checkpoint-Restart is specified.
• The criteria for an incore sort are not met. (See the discussion of incore sorts in
“Chapter 14 Performance Considerations”.)
• SUM, DUPKEYS, OUTREC or OUTFIL is used.
• SORTOUT is a VSAM data set.

Syncsort MFX Programmer’s Guide 4–11

 Note: Sort applications that use SUM, DUPKEYS, OUTREC, OUTFIL or VSAM
SORTOUT and do not provide JCL SORTWORKs may have DYNALLOC automat-
ically enabled. This will allow the completion of a sort that would have terminated
for lack of required SORTWORK space.

SYSIN DD Statement
The data set defined by the SYSIN DD statement contains MFX control
statements. The SYSIN DD statement is required in order to initiate the sort/
merge through job control language.

//SYSIN DD *
SORT FIELDS=(5,3,CH,A)
OMIT COND=(12,6,PD,EQ,0)
END
/*

Figure 255. Sample SYSIN DD Statement

$ORTPARM DD Statement
The data set defined by the $ORTPARM DD statement may contain PARM
parameters and any of the sort control statements.
Parameters and control statements passed via the $ORTPARM DD statement
generally override all others passed, whether the sort/merge is called from a
program or initiated through job control language.
The $ORTPARM DD record format must be F or FB, and the record length must be
80 bytes. Labels are not allowed on $ORTPARM card images. Leading blanks are
not required on a PARM card image, but at least one leading blank must precede a
sort control statement keyword.
The $ORTPARM data sets must be formatted in accordance with the following
rules:
• PARM specifications included in the $ORTPARM data sets must be specified
before any sort control statement specifications.
• PARMS must be specified without the keyword PARM= and without quotation
marks.
• A comma in columns 2-70 of a PARM card image followed by a blank, or a
comma alone in column 71, may be used to indicate that the next record is part
of the current statement. However, if the PARM specification is present through

4–12 Syncsort MFX Programmer’s Guide

 column 71, a continuation character must be specified in column 72 to indicate
continuation.
• Comments may be included on $ORTPARM card images provided there is a
blank between the last PARM specification and the comment. You may continue
a comment by placing a continuation character in column 72 if there are no
additional PARMs. In this case, the entire next card image will be considered a
comment. If additional PARMs will follow the comment, you may continue that
comment by coding an asterisk (*) in column 1 of the next card image.
Note: Refer to “Chapter 2 MFX Control Statements” for additional formatting
requirements.
The following example of a $ORTPARM data set illustrates the conventions for
defining the $ORTPARM data set.

//$ORTPARM DD *
BMSG,STOPAFT=500,
EQUALS
SORT FIELDS=(1,8,PD,A)

Figure 256. Sample $ORTPARM DD Statement

The $ORTPARM data set in the previous example overrides the options set in the
associated invoking program (or job control stream) to sort 500 records from the
input file. These will be the first 500 records that meet whatever criteria have been
set by the original application (which might include, for example, the INCLUDE/
OMIT control statement). BMSG turns on the WERnnnB message set, so that the
processing accorded these 500 records is fully documented. EQUALS preserves the
order of equal-keyed records from input to output.

//$ORTPARM DD *
BMSG,STOPAFT=500,
EQUALS
SUM FIELDS=(12,4,30,8,38,8),
FORMAT=PD
SORT FIELDS=(1,8,PD,A)

Figure 257. Sample $ORTPARM DD Statement

The preceding example illustrates how to include control statements more than 80
bytes long; continuation card images are indicated by a blank field following an
operand-comma combination.

Syncsort MFX Programmer’s Guide 4–13

 //SYSIN DD *
OUTFIL FILES=(1,2,3),
.
.
.
//$ORTPARM DD *
OUTFIL FILES=(3,4,5),
.
.
.

Figure 258. Sample $ORTPARM DD Statement

In this example, the OUTFIL control statement in $ORTPARM overrides the

OUTFIL control statement in SYSIN for file 3, and adds OUTFIL specifications for
files 4 and 5.

$ORTPARM Processing for Century Window COBOL Applications

The $ORTPARM DD facility is particularly useful for COBOL sorts requiring
century window processing of year data with MFX’s year data formats. The year
data formats are not supported by COBOL. Therefore, when a data format
specification needs to be changed for century window processing, it is necessary to
override SORT control statements generated by COBOL. The override can be
accomplished with a $ORTPARM DD statement. The following example shows a
$ORTPARM DD used for this purpose.

//$ORTPARM DD *
SORT FIELDS=(10,2,Y2Z,A),CENTWIN=1980

Figure 259. Sample $ORTPARM DD Statement for Century Window Processing

In this example, the 2-digit year field (10,2) will have century window processing
applied to it via the Y2Z year data format and the CENTWIN option.
As described in the previous section, multiple sort invocations by the same COBOL
program would require multiple $ORTPARM DD statements, each with the
FREE=CLOSE parameter.

$ORTPARM DD Processing for Multiple Sort Invocations

When MFX is to be invoked more than once in the same job step, you may need
different $ORTPARM DD control data sets for each invocation. For multiple control
data sets, define each one in the JCL stream, in the desired order, as a disk data set
(or partitioned data set member) with the FREE=CLOSE parameter added.
FREE=CLOSE will cause the first sort $ORTPARM data set to be dynamically
deallocated by the first sort execution, and so forth for each sort execution. The
following example shows sample JCL with two $ORTPARM DD statements:

4–14 Syncsort MFX Programmer’s Guide

 //$ORTPARM DD DSN=SORT.OPTIONS(SORT1),DISP=SHR,FREE=CLOSE
//* WILL BE USED BY FIRST SORT EXECUTION
//$ORTPARM DD DSN=SORT.OPTIONS(SORT2),DISP=SHR,FREE=CLOSE
//* WILL BE USED BY SECOND SORT EXECUTION
.
.
//$ORTPARM DD DSN=SORT.OPTIONS(SORTn),DISP=SHR,FREE=CLOSE
//* WILL BE USED BY THE nTH SORT EXECUTION

Figure 260. Sample Multiple $ORTPARM DD Statements

Processing will proceed from top to bottom of this $ORTPARM data set list. This
sequence must be maintained in the JCL so that the multiple sorts can read the
$ORTPARM data sets in the correct order.
Multiple $ORTPARM data sets are available only in a JES2 environment. JES3
does not support the specification of multiple DD statements for the same
DDNAME.

SORTCKPT DD Statement
This DD statement is only used when the CKPT/CHKPT option is set on the
SORT/MERGE control statement, requesting the Checkpoint-Restart feature. See
“The Coding and Use of Checkpoint-Restart” on page 14-7 for an explanation of this
feature.

For Exit Routines that Require Link-editing at Execution

Time
The following DD statements are required whenever an exit routine is to be link-
edited at execution time.

SORTMODS DD Statement
The partitioned data set defined must be large enough to contain all the exit
routines entered in SYSIN. For exits not entered in SYSIN, it is necessary to supply
DD statements defining the libraries in which the routines reside.

//SORTMODS DD SPACE=(CYL,(2,,4)),UNIT=SYSDA

Figure 261. Sample SORTMODS DD Statement

Syncsort MFX Programmer’s Guide 4–15

SYSLIN DD Statement
The SYSLIN DD statement defines the temporary data set that will contain the
linkage editor control statements created by MFX for the exit routine(s).

//SYSLIN DD DSN=&&TEMP,UNIT=SYSDA,SPACE=(TRK,1)

Figure 262. Sample SYSLIN DD Statement

SYSLMOD DD Statement
The SYSLMOD DD statement defines the temporary data set that will contain the
link-edited exit module(s).

//SYSLMOD DD DSN=&&TEMP2,UNIT=SYSDA,
// SPACE=(TRK,(10,5,2))

Figure 263. Sample SYSLMOD DD Statement

SYSPRINT DD Statement
The SYSPRINT DD statement defines the message data set for the link-editing of
sort exits.

//SYSPRINT DD SYSOUT=A

Figure 264. Sample SYSPRINT DD Statement

STDOUT DD Statement
The STDOUT DD statement defines the informational message data set used by
Java during the creation of PDF, RTF and HTML files via the OUTFIL OUTPUT
facility.

//STDOUT DD SYSOUT=*

Figure 265. Sample STDOUT DD Statement

STDERR DD Statement
The STDERR DD statement defines the error message data set used by Java
during the creation of PDF, RTF and HTML files via the OUTFIL OUTPUT facility.

//STDERR DD SYSOUT=*

Figure 266. Sample STDERR DD Statement

4–16 Syncsort MFX Programmer’s Guide

DD Statements for MAXSORT, PARASORT, DB2 Query and
MULTIIN Support
The MAXSORT technique is initiated by means of the MAXSORT PARM, and
utilizes additional MAXSORT DD statements (SORTBKPT, SORTOU00,
SORTOUnn) and PARMs. With MAXSORT, SORTWK files must be allocated to
disk devices. This technique is strongly recommended for very large sorting
applications in a limited disk work space environment. See “Chapter 9 MAXSORT”.
The PARASORT technique is initiated by means of the PARASORT PARM and
utilizes additional PARASORT DD statements (SORTPAR1, SORTPAR2,
SORTPAR3, SORTPAR4). PARASORT requires disk SORTWK devices. This
technique can improve the elapsed time of sorting applications that have multi-
volume tape SORTIN data sets. See “Chapter 10 PARASORT”.
The DB2 Query Support technique is initiated by means of the DB2 Query Support
PARM and utilizes the DB2 Query Support DD statement SORTDBIN. This
technique allows DB2 data to be passed directly into a SORT or COPY operation,
without the use of setup steps or the need for user-written E15 exits. See “Chapter
11 MFX DB2 Query Support”.
The MULTIIN facility is initiated by means of the MULTIIN PARM and utilizes
SORTMInn DD statements. This facility allows the combination of VSAM and non-
VSAM files as input to a SORT or COPY operation. See “Chapter 12 Multiple Input
Files”.

Sample JCL/Control Statement Streams

The sample JCL/control statement streams in this section illustrate how to specify
sort, merge and copy applications with and without exit routines. An example
illustrating multiple output is also included. Refer to “Chapter 3 How to Use the
MFX Data Utility Features” for comprehensive examples illustrating the data
utility and report writing features. Examples of how to invoke MFX from a
program, COBOL exit routines, MAXSORTs, and PARASORTs are provided in the
appropriate chapters.

Syncsort MFX Programmer’s Guide 4–17

Sorts without Exit Routines
Example 1

//SORTOMIT JOB 1
//SORT1 EXEC PGM=SYNCSORT,PARM='STOPAFT=1000' 2
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 3
//SYSOUT DD SYSOUT=A 4
//SORTIN DD DSN=INPUT,UNIT=3490, 5
// VOL=SER=012345,DISP=(OLD,KEEP),
// DCB=(LRECL=100,RECFM=FB,
// BLKSIZE=32700),LABEL=(1,SL)
//SORTOUT DD DSN=OUTPUT,VOL=SER=543210, 6
// UNIT=3490,DISP=(NEW,KEEP),
// DCB=(LRECL=100,RECFM=FB,
// BLKSIZE=0),LABEL=(1,SL)
//SORTWK01 DD SPACE=(CYL,(20)),UNIT=SYSDA 7
//SORTWK02 DD SPACE=(CYL,(20)),UNIT=SYSDA
//SORTWK03 DD SPACE=(CYL,(20)),UNIT=SYSDA
//SORTWK04 DD SPACE=(CYL,(20)),UNIT=SYSDA
//SORTWK05 DD SPACE=(CYL,(20)),UNIT=SYSDA
//SYSIN DD * 8
SORT FIELDS=(1,8,CH,A) 9
OMIT COND=(1,8,CH,EQ,C'JOHN DOE') 10
END 11
/* 12

Figure 267. Sample JCL/Control Stream (1)

1. The JOB statement gives SORTOMIT as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed. The

STOPAFT PARM instructs MFX to terminate after sorting 1,000 records.

3. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP indicates that this library may be shared.

4. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

5. The SORTIN DD statement gives INPUT as the input data set name, specifies a 3490
tape unit with the volume serial number 012345. The data set is already in existence.

The DCB parameter shows an LRECL of 100 bytes, a fixed blocked RECFM,
and a 32700-byte BLKSIZE. The LABEL parameter shows that INPUT is the
first data set on the tape, and that it has a standard label.

6. The SORTOUT DD statement gives OUTPUT as the output data set name, and
specifies a 3490 tape unit with the volume serial number 543210. The data set is not in
existence yet.

4–18 Syncsort MFX Programmer’s Guide

 The DCB parameter for SORTOUT specifies the same LRECL and RECFM as
SORTIN. The BLKSIZE will be selected by System Determined BLKSIZE
(SDB) if active or by MFX if SDB is not active.

7. The five SORTWKxx DD statements reserve space on direct access devices for
intermediate storage. Twenty cylinders are allocated for each of the five SORTWKxx
data sets.

8. The SYSIN DD * statement marks the beginning of the system input stream that
includes the sort control statements.

9. The SORT control statement specifies that one control field will be sorted on. It begins
on byte 1 of the record, is 8 bytes long, contains character data, and is to be sorted in
ascending order.

10. The OMIT control statement eliminates any record with JOHN DOE in its first eight
bytes (i.e., in the sort control key). JOHN DOE records are not sorted and are not
included in the STOPAFT figure. The EXEC statement’s STOPAFT PARM terminates
the sort after 1,000 (non-JOHN DOE) records have been put into the proper sequence.

11. The END control statement marks the end of the control statements.

12. The delimiter statement marks the end of the SYSIN input stream.

Example 2

//SUMSORT JOB 1
// EXEC PGM=SYNCSORT,PARM='EQUALS' 2
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 3
//SYSOUT DD SYSOUT=A 4
//SORTIN DD DSN=FEB92,EMPLOYEE.MASTER, 5
// UNIT=3490,VOL=SER=135790,
// DISP=(OLD,KEEP)
// DD DSN=FEB92.EMPLOYEE.UPDATE,
// UNIT=3490,VOL=SER=999999,
// DISP=(OLD,KEEP)
//SORTOUT DD DSN=MAR92.EMPLOYEE.MASTER, 6
// UNIT=3490,VOL=SER=246809,
// DISP=(NEW,KEEP)
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,20) 7
//SYSIN DD * 8
SORT FIELDS=(1,9,ZD,A,10,2,BI,A) 9
SUM FIELDS=(12,4,PD) 10
/* 11

Figure 268. Sample JCL/Control Stream (2)

Syncsort MFX Programmer’s Guide 4–19

 1. The JOB statement gives SUMSORT as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed. The

EQUALS PARM interacts with the SUM control statement to preserve the first of a
series of equal-keyed records.

3. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP shows the library may be shared.

4. The SYSOUT DD statement assigns the MFX messages to the output device associated
with class A.

5. The SORTIN DD statements define the input to be copied: two concatenated data sets
— FEB92.EMPLOYEE.MASTER and FEB92.EMPLOYEE.UPDATE. They are found
on standard labeled 3490 tape units (volume serial numbers 135790 and 999999,
respectively). These data sets are already in existence.

6. The SORTOUT DD statement gives MAR92.EMPLOYEE.MASTER as the output data

set name and specifies a 3490 tape unit with the volume serial number 246809. The
data set is not in existence yet.

The DCB RECFM and LRECL parameters for SORTOUT default to that of the
first SORTIN file. The BLKSIZE will be selected by System Determined
BLKSIZE (SDB) if active or by MFX if SDB is not active.

7. The SORTWK01 DD statement reserves space on a direct access device for

intermediate storage. Twenty cylinders are allocated. Intermediate storage must be
provided whenever the SUM control statement is used with a sort.

8. The SYSIN DD statement marks the beginning of the system input stream that
includes the sort control statements.

9. The SORT control statement specifies that two control fields will be sorted on. The
major control field begins on byte 1 of the record, is 9 bytes long, contains zoned decimal
data, and is to be sorted in ascending numerical order. The second, less significant,
control field is found in the next two bytes of the record (bytes 10 and 11), is in
(unsigned) binary format, and is to be sorted in ascending order.

10. Whenever two records have equal control fields, the sort will attempt to sum them. If
the result of summing the packed decimal data found in the 4-byte field beginning at
byte 12 can be contained in four bytes, one of the two records will be retained, the sum
stored in bytes 12-15, and the other record will be deleted. The EQUALS PARM
guarantees that the first of the two records will be preserved; thus, if a record from the
FEB92.EMPLOYEE.MASTER file has the same key as one from the
FEB92.EMPLOYEE.UPDATE file, it is the master record which is retained in the
output file, containing their sum.

11. The delimiter statement marks the end of the SYSIN input stream.

4–20 Syncsort MFX Programmer’s Guide

Example 3

//SORTSKIP JOB 1
// EXEC PGM=SYNCSORT 2
//$ORTPARM DD * 3
STOPAFT=100
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 4
//SYSOUT DD SYSOUT=A 5
//SORTIN DD DSN=EXPORT.SHIPPING.VOL6, 6
// UNIT=TAPE,VOL=SER=112233,
// DISP=(OLD,KEEP)
//SORTOUT DD DSN=RECENT.MAJOR.EXPORTS, 7
// UNIT=TAPE,VOL=SER=332211,
// DISP=(NEW,KEEP)
//SORTWK01 DD SPACE=(CYL,20),UNIT=SYSDA 8
//SORTWK02 DD SPACE=(CYL,20),UNIT=SYSDA
//SORTWK03 DD SPACE=(CYL,20),UNIT=SYSDA
//SYSIN DD * 9
SORT FIELDS=(19,5,CH,A), 10
EQUALS,SKIPREC=1000
INCLUDE COND=(37,4,BI,GE,X'50') 11
/* 12

Figure 269. Sample JCL/Control Stream (3)

1. The JOB statement gives SORTSKIP as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed.

3. The $ORTPARM DD statement is used here to initiate a test run of the SORTSKIP job
by supplying the STOPAFT PARM to MFX. It instructs MFX to terminate after sorting
the first 100 of the records INCLUDE selects from the SKIPREC-edited input file.

4. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP indicates that this library may be shared.

5. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

6. The SORTIN DD statement gives EXPORT.SHIPPING.VOL6 as the input data set

name. It is found on a standard labeled tape having the volume serial number 112233.
This data set is already in existence.

7. The SORTOUT DD statement assigns the RECENT.MAJOR.EXPORTS data set name

to the output file, and specifies a tape unit with the volume serial number 332211. This
data set is not yet in existence. The DCB RECFM and LRECL parameters for
SORTOUT default to those of the first SORTIN file. The BLKSIZE will be selected by
System Determined BLKSIZE (SDB) if active or by MFX if SDB is not active.

Syncsort MFX Programmer’s Guide 4–21

 8. The three SORTWKxx DD statements reserve space on direct access devices for
intermediate storage. Twenty cylinders are allocated for each SORTWK data set.

9. The SYSIN DD statement marks the beginning of the system input stream that
includes the sort control statements.

10. The SORT control statement specifies that one control field will be sorted on. It begins
on byte 19 of the record, is 5 bytes long, contains character data, and is to be sorted
according to ascending order. The EQUALS parameter preserves the SORTIN order of
records with identical data in these five bytes. The SKIPREC parameter eliminates the
first 1,000 records of the SORTIN file from consideration; these records are eliminated
before the INCLUDE statement takes effect.

11. The INCLUDE statement compares the 4 bytes beginning with byte 37 of the record to
the hexadecimal literal, which will be padded on the right with binary zeros to the
indicated (4 byte) length. The record is eliminated from the sort unless the binary data
in that field is at least as great as the padded constant. The INCLUDE/OMIT
statement takes effect after SKIPREC but before STOPAFT.

12. The delimiter statement marks the end of the SYSIN input stream.

A Merge without Exit Routines

Example 4

//EDITMERG JOB 1
//MERGE1 EXEC PGM=SYNCSORT 2
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 3
//SYSOUT DD SYSOUT=A 4
//SORTIN08 DD DSN=SALES91,UNIT=TAPE, 5
// VOL=SER=123456,DISP=(OLD,KEEP)
//SORTIN12 DD DSN=SALES92,UNIT=TAPE,
// VOL=SER=654321,DISP=(OLD,KEEP)
//SORTIN03 DD DSN=SALES93,UNIT=3390,
// VOL=SER=DISK11,DISP=SHR
//SORTOUT DD DSN=SALES.PATTERN,UNIT=3390, 6
// VOL=SER=DISK08,DISP=(NEW,KEEP),
// SPACE=(CYL,5),
// DCB=(LRECL=20,RECFM=VB,
// BLKSIZE=27980)
//SYSIN DD * 7
MERGE FIELDS=(5,4,ZD,A) 8
RECORD TYPE=V,LENGTH=(100,,20) 9
INREC FIELDS=(1,8,29,6,12,6) 10
/* 11

Figure 270. Sample JCL/Control Stream (4)

4–22 Syncsort MFX Programmer’s Guide

 1. The JOB statement gives EDITMERG as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed.

3. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP shows the library may be shared.

4. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

5. Three data sets are to be merged: SALES91, SALES92 and SALES93. SALES91 and
SALES92 are found on standard labeled tapes with the volume serial numbers 123456
and 654321, respectively. The DD statement for SALES93 specifies a 3390 disk device
with the volume serial number DISK11. These three data sets are already in existence,
and the disk data set SALES93 may be shared. They are assigned distinct SORTINnn
numbers, as required.

6. The SORTOUT DD statement assigns the name SALES.PATTERN to the output data
set and specifies a 3390 disk device with the volume serial number DISK08. Five
cylinders of primary space have been allocated on this volume. The data set does not yet
exist. DCB parameters are provided, preventing them from defaulting to those of the
SORTIN08 file.

7. The SYSIN DD statement marks the beginning of the system input stream that
includes the sort control statements.

8. The MERGE control statement specifies one control field. It begins on byte 5 (the first
data byte of the record since TYPE=V is specified on the RECORD statement) and is 4
bytes long. This field contains zoned decimal data and is to be merged in ascending
order.

9. The RECORD statement indicates that variable-length records are being merged and
indicates the record length at various processing stages. The maximum input record
length is specified as 100 bytes. Since there is no E15, the post-E15 length value is not
coded and so defaults to this figure. The INREC statement reduces this maximum
record length to just 20 bytes.

10. According to the RECORD control statement, the input record may be 100 bytes long.
The INREC statement reduces each record to the 20 bytes crucial to this application:
the 4-byte RDW and 4-byte merge control field (i.e., the first 8 bytes of the record), the
6-byte field beginning at byte 29 (the 25th data byte) and the 6-byte field beginning at
byte 12 (the 8th data byte). As required, the RDW remains in the first four bytes. The
records to be merged are no more than 20 bytes long and contain three fields following
the RDW.

11. The delimiter statement marks the end of the SYSIN input stream.

Syncsort MFX Programmer’s Guide 4–23

A Copy without Exit Routines

Example 5

//COPYNYC JOB 1
//COPY1 EXEC PGM=SYNCSORT 2
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 3
//SYSOUT DD SYSOUT=A 4
//SORTIN DD DSN=USA.OUTLETS,UNIT=TAPE, 5
// VOL=SER=149200,DISP=(OLD,KEEP)
//SORTOUT DD DSN=NYC.OUTLETS,UNIT=3390, 6
// VOL=SER=DISK08,SPACE=(CYL,5),
// DISP=(NEW,KEEP)
//SYSIN DD * 7
SORT FIELDS=COPY 8
INCLUDE COND=(56,3,CH,EQ,C'NYC') 9
/* 10

Figure 271. Sample JCL/Control Stream (5)

1. The JOB statement gives COPYNYC as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed.

3. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP shows the library may be shared.

4. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

5. The SORTIN DD statement indicates the file to be copied. The data set name is
USA.OUTLETS, and it is found on the standard labeled tape with the volume serial
number 149200. The data set is already in existence.

6. The SORTOUT DD statement names the copied file NYC.OUTLETS, and specifies a
3390 disk device with the volume serial number of DISK08. Five cylinders of primary
space have been allocated on this volume. The data set does not yet exist, but is to be
kept whether or not the job terminates normally. The DCB RECFM and LRECL
parameters for SORTOUT default to that of the first SORTIN file. The BLKSIZE will
be selected by System Determined BLKSIZE (SDB) if active or by MFX if SDB is not
active.

7. The SYSIN DD statement marks the beginning of the system input stream that
includes the sort control statements.

8. The FIELDS parameter specifies a copy application. This could have been coded as
MERGE FIELDS=COPY without affecting program execution.

4–24 Syncsort MFX Programmer’s Guide

 9. The INCLUDE control statement edits the USA.OUTLETS input file, eliminating all
records which do not have the character string NYC in bytes 56-58. Only 'NYC' records
will be copied.

10. The delimiter statement marks the end of the SYSIN input stream.

A Sort with an Exit Routine Already Link-edited

Example 6

//ONE#EXIT JOB 1
//STEP1 EXEC PGM=SYNCSORT,PARM='MSG=SC' 2
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 3
//SYSOUT DD SYSOUT=A 4
//MODLIB DD DSN=EXIT.E15,DISP=SHR 5
//SORTIN DD DSN=INPUT,UNIT=3390, 6
// VOL=SER=ABCDEF,DISP=(SHR)
//SORTOUT DD DSN=OUTPUT,UNIT=3390, 7
// VOL=SER=GHIJKL,SPACE=(CYL,10)
// DISP=(NEW,KEEP,DELETE)
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,20) 8
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,20)
//SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,15)
//SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,15)
//SYSIN DD * 9
SORT FIELDS=(10,25,CH,A,40,10,ZD,D), 10
FILSZ=9000
RECORD TYPE=V,LENGTH=(1024,,,44,192) 11
MODS E15=(E15,600,MODLIB,N) 12
END 13
/* 14

Figure 272. Sample JCL/Control Stream (6)

1. The JOB statement gives ONE#EXIT as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed. The MSG
PARM option requests that all messages be routed to the SYSOUT DD statement but
only critical messages be routed to the console.

3. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP shows the library may be shared.

4. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

5. The MODLIB DD statement defines the library in which the exit routine resides;
MODLIB is referenced in the MODS control statement. The data set name of the
library is EXIT.E15, and the DISP shows that the library may be shared.

Syncsort MFX Programmer’s Guide 4–25

 6. The SORTIN DD statement gives INPUT as the input data set name and specifies a
3390 disk with the volume serial number ABCDEF. The DISP parameter indicates that
the data set is already in existence and may be shared.

7. The SORTOUT DD statement gives OUTPUT as the output data set name and specifies
a 3390 disk with the volume serial number GHIJKL. Ten cylinders of primary space
have been allocated on this volume. The DISP parameter shows that this data set is not
yet in existence.

8. The four SORTWK statements reserve space on four temporary data sets for
intermediate storage. Twenty cylinders are to be reserved on the first two data sets,
fifteen on the second two data sets.

9. The SYSIN DD statement marks the beginning of the input stream that includes the
sort control statements.

10. The SORT control statement specifies two sort control fields. The first begins on byte 10
(data byte 6) of the record, is 25 bytes long, contains character data, and is to be sorted
in ascending order. The second control field begins on byte 40 (data byte 36) of the
record, is 10 bytes long, has zoned decimal data, and is to be sorted in descending order.
FILSZ instructs MFX to terminate abnormally unless the post-E15 file contains exactly
9,000 records.

11. The RECORD control statement shows that variable-length records are being sorted.
The first LENGTH value reports that the maximum length of records in the SORTIN
data set is 1024 bytes. The comma coded for the second LENGTH value shows that this
maximum length is not altered by the exit routine. The comma coded for the third
LENGTH value shows that this maximum length is not affected by an E35 or the
INREC/OUTREC statements. The fourth LENGTH value shows that the smallest
record in the input data set is 44 bytes long. The fifth LENGTH value shows that the
record length that occurs most frequently in SORTIN is 192 bytes. (This value will be
used to determine segment size.)

12. The MODS control statement states that the exit-type is E15. The name of the actual
exit routine included at this exit is also E15. The routine requires 600 bytes of memory
and resides in a library defined on the MODLIB DD statement. Finally, the N indicates
that link-editing of the routine has already been performed.

13. The END control statement marks the end of the control statements.

14. The delimiter statement marks the end of the SYSIN input stream.

4–26 Syncsort MFX Programmer’s Guide

A Sort with an Exit Routine to be Link-edited
Example 7

//LINKEXIT JOB 1
//STEP EXEC PGM=SYNCSORT 2
//SYSOUT DD SYSOUT=A 3
//SORTIN DD DSN=IN.FILE.JANUARY, 4
// UNIT=TAPE,VOL=SER=135790,
// DISP=OLD,DELETE),
// DCB=(LRECL=200,RECFM=FB,
// BLKSIZE=4000),LABEL=(2,SL)
//SORTOUT DD DSN=OUT.FILE.FEBRUARY, 5
// UNIT=TAPE,VOL=SER=097863,
// DISP=(NEW,KEEP),LABEL=(1,SL)
//SORTMODS DD DSN=A.PART.DATA.SET,DISP=OLD 6
//MODLIB DD DSN=EXIT.NO.ONE,DISP=SHR 7
//SYSLMOD DD DSN=&&LINK,UNIT=SYSDA, 8
// SPACE=(CYL,(1,1,1))
//SYSLIN DD DSN=&&TEMP,UNIT=SYSSQ, 9
// SPACE=(TRK,1)
//SYSPRINT DD SYSOUT=A 10
//SYSIN DD * 11
SORT FIELDS=(20,30,CH,A), 12
DYNALLOC=(SYSDA,6)
RECORD TYPE=F,LENGTH=200 13
MODS E15=(EXIT1,600,MODLIB,N), 14
E35=(EXIT2,500,SYSIN)
SUM FIELDS=(1,10,ZD) TOTAL BALANCE 15
END ACCOUNTS FOR JANUARY BEGIN FEBRUARY 16
.
.
.
Object deck EXIT2 for E35 exit routine 17
.
.
.
/* 18

Figure 273. Sample JCL/Control Stream (7)

1. The JOB statement gives LINKEXIT as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed.

3. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

4. The SORTIN DD statement gives IN.FILE.JANUARY as the input data set name, and
specifies a tape unit with the volume serial number 135790. The DISP parameter
shows that the data set is already in existence.

Syncsort MFX Programmer’s Guide 4–27

 The DCB parameter shows an LRECL of 200 bytes, a fixed blocked RECFM,
and a 4000-byte BLKSIZE. The LABEL parameter shows that
IN.FILE.JANUARY is the second data set on the tape, and that it has a
standard label.

5. The SORTOUT DD statement gives OUT.FILE.FEBRUARY as the output data set

name, and specifies a tape unit with the volume serial number 097863. The DISP
parameter shows that the data set is not in existence yet.

The DCB RECFM and LRECL parameters for SORTOUT default to that of the
first SORTIN file. The BLKSIZE will be selected by System Determined
BLKSIZE (SDB) if active or by MFX if SDB is not active. The LABEL
parameter shows that OUT.FILE.FEBRUARY is to be the first data set on the
tape, and will have a standard label.

6. The SORTMODS DD statement defines the partitioned data set that will contain the
exit routine object module that has not been link-edited and is being included in the
SYSIN data stream. The DISP shows the data set may not be shared.

7. The MODLIB DD statement defines the partitioned data set in which the already link-
edited exit routine resides. (Note MODLIB is referenced on the MODS control
statement.) The data set name of the exit library is EXIT.NO.ONE. The DISP shows the
data set may be shared.

8. The SYSLMOD DD statement defines a temporary data set called &&LINK that will
contain the exit routine after it has been link-edited. A direct access device will be used
with 1 cylinder reserved for primary space allocation, 1 cylinder for secondary space
allocation, and 1 directory block.

9. The SYSLIN DD statement defines the temporary data set that will contain the linkage
editor control statements that MFX will use when link-editing the exit. The name of
this data set is &&TEMP. It is to be on any sequential-access device with 1 track
reserved if the data set is allocated to disk.

10. The SYSPRINT DD statement defines the data set on which the linkage editor will
write its messages. Whatever device is assigned to SYSOUT=A will be used.

11. The SYSIN DD statement marks the beginning of the input stream that includes the
sort control statements and also the object deck of the exit routine to be link-edited.

12. The SORT control statement shows that one control field will be sorted on. It begins on
byte 20 of the record, is 30 bytes long, contains character data, and is to be sorted
according to ascending order.

The DYNALLOC parameter specifies that 6 direct access areas are to be

reserved for sortwork data sets.

13. The RECORD control statement shows that fixed-length records are being sorted. The
LENGTH parameter gives 200 bytes as the length of the records at input time, and, by

4–28 Syncsort MFX Programmer’s Guide

 not specifying values for l2 and l3, implicitly states that the length of these records will
not be changed during the sort.

14. The MODS control statement shows that the first exit-type is E15. The name of the
routine for this exit is EXIT1. It will take 600 bytes in main storage, resides in a library
defined on the MODLIB DD statement, and has already been link-edited.

The second exit-type is E35. The name of the routine for the exit is EXIT2, and
it will take 500 bytes in main storage. The object deck for the routine is to be
included in the SYSIN portion of the job stream, and, because of the absence of
a letter in the last subparameter position for this group, the sort assumes that
the routine requires link-editing and will be link-edited together with any other
routines for this phase.

15. The SUM control statement’s FIELDS parameter identifies one summed field. It begins
on byte 1 of the record, is 10 bytes long, and has zoned decimal data. The rest of the
statement is a comment.

16. The END control statement marks the end of the control statements and also contains a
comment.

17. The EXIT2 object deck to be link-edited is included after the END statement in the
SYSIN stream.

18. The delimiter statement marks the end of the SYSIN input stream for the sort.

Multiple Output Files

Example 8

Syncsort MFX Programmer’s Guide 4–29

 //MULTOUT JOB 1
// EXEC PGM=SYNCSORT 2
//STEPLIB DD DSN=SORT.RESI.DENCE,DISP=SHR 3
//SYSOUT DD SYSOUT=A 4
//SORTIN DD DSN=SALES.RECORDS, 5
// VOL=SER=DISK1,DISP=SHR
//SORTOUT DD DSN=SORTED.SALES.RECORDS, 6
// UNIT=TAPE,VOL=SER=112233,
// DISP=(NEW,KEEP)
//SORTOFDS DD DSN=DOMESTIC.SALES.RECORDS, 7
// VOL=SER=DISK8,DISP=(NEW,KEEP),
// SPACE=(CYL,40),UNIT=SYSDA
//SORTWK01 DD SPACE=(CYL,20),UNIT=SYSDA 8
//SORTWK02 DD SPACE=(CYL,20),UNIT=SYSDA
//SORTWK03 DD SPACE=(CYL,20),UNIT=SYSDA
//SYSIN DD * 9
SORT FIELDS=(10,12,BI,A) 10
OUTFIL FILES=OUT,INCLUDE=ALL 11
OUTFIL FILES=DS,OMIT=(62,3,CH,NE,C'USA') 12
/* 13

Figure 274. Sample JCL/Control Stream (8)

1. The JOB statement gives MULTOUT as the jobname.

2. The EXEC statement identifies SYNCSORT as the program to be executed.

3. The STEPLIB DD statement instructs the system to look for MFX in the library named
SORT.RESI.DENCE. The DISP parameter shows that the library may be shared.

4. The SYSOUT DD statement assigns the MFX messages to the output device associated
with SYSOUT class A.

5. The SORTIN DD statement gives SALES.RECORDS as the input data set name, and
specifies a disk with the volume serial DISK1. The DISP parameter indicates that the
data set is already in existence and may be shared.

6. The SORTOUT DD statement names one of the output sorted files

SORTED.SALES.RECORDS, and specifies a tape device with volume serial number
112233 for storage. The DISP parameter indicates that the data set does not yet exist,
but it is to be kept whether or not the job terminates normally.

7. The SORTOFDS DD statement names a second sorted output file

DOMESTIC.SALES.RECORDS, and specifies a disk device with volume serial number
DISK8 for storage. Forty cylinders of space have been allocated on this volume. The
DISP parameter indicates that the data set does not yet exist, but is to be kept whether
or not the job terminates normally.

4–30 Syncsort MFX Programmer’s Guide

 8. The three SORTWK DD statements reserve space on direct access devices for
intermediate storage. Twenty cylinders are allocated for each of the three SORTWK
data sets.

9. The SYSIN DD statement marks the beginning of the input stream that includes the
sort control statements.

10. The SORT control statement specifies that one control field will be sorted on. It begins
on byte 10 of the record, is 12 bytes long, contains unsigned binary (BI) data and is to be
sorted according to ascending order.

11. The first OUTFIL control statement is associated with the SORTOUT DD statement.
The INCLUDE parameter specifies that all input records are to be included in this
output file.

12. The second OUTFIL control statement is associated with the SORTOFDS DD
statement. The OMIT parameter specifies that records which do not contain “USA” in
bytes 62, 63 and 64 are not to be included in this file.

13. The delimiter statement marks the end of the SYSIN input stream.

Syncsort MFX Programmer’s Guide 4–31

4–32 Syncsort MFX Programmer’s Guide
Chapter 5 PARM Options

PARM options can be specified to provide processing information and to override

installation defaults for JCL-initiated and program-invoked applications.
For a JCL-initiated application, specify the PARM option(s) on the EXEC
statement as follows:

PARM='option,...'

Figure 275. PARM Parameter Format

For a program-invoked application, specify the PARM option(s) in a $ORTPARM

DD data set. Omit the keyword PARM= and the single quotes. PARM options for a
JCL-initiated application can also be specified in a $ORTPARM data set.

Precedence Rules
There are three ways in which options can be specified, though not all options can
be specified in all three ways:
• As an installation specification
• As a PARM specification
• As a SORT/MERGE control statement specification.
Note that there are six options that can be specified as a PARM option or a SORT/
MERGE option. They are: CENTWIN, DYNALLOC, EQUALS/NOEQUALS, FILSZ,
SKIPREC, and STOPAFT.
When an option is specified in more than one way, the following precedence rules
apply:

Syncsort MFX Programmer’s Guide 5–1

 • A SORT/MERGE or PARM specification overrides an installation specification.
• A PARM specification overrides a SORT/MERGE specification except for
EQUALS/NOEQUALS.

PARM Option Summary Chart

The chart on the following pages lists the PARM options. Underscored PARM
options are delivered defaults which may have been altered at installation time.
PARM Option Name Description
BALANCE Balances importance of CPU time, elapsed time, and I/O activity
for best overall sort performance. See CPU, ELAP, and IO. Note
that these options and BALANCE are all mutually exclusive.
BMSG Produces WERnnB messages.
CENTWIN= 0 /s /f Generates a sliding (s) or fixed (f) 100-year window that deter-
mines the century to which 2-digit year data belongs. Ensures
that such data is processed correctly as a 4-digit year by SORT/
MERGE and INCLUDE OMIT. Also enables OUTREC processing
to output a 4-digit year (yyyy) from 2-digit year input (yy).
CMP= CPD/CLC CMP=CPD improves performance.
COMMAREA/NOCOMMAREA Provides a communication area between exit programs.
CORE/SIZE=n Changes the amount of memory in which sort/merge can run.
CPU Minimizes CPU time at expense of other performance measures.
See BALANCE, ELAP, and IO. Note that these options and CPU
are all mutually exclusive.
DEBUG Provides an MFX SNAP dump in the event of a critical error.
DIAG Provides diagnostic information for certain error conditions.
DYNALLOC Requests the dynamic allocation of work data sets.
E15/E35=COB Indicates a COBOL exit.
ELAP Minimizes elapsed time at expense of other performance
measures. See BALANCE, CPU, and IO. Note that these options
and ELAP are all mutually exclusive.
EQUALS/ NOEQUALS EQUALS acts to preserve the order of equal-keyed records. It is
not available with PARASORT.
EXTCOUNT Enables special processing for applications with record counts
that exceed MFX’s default internal limit. .
FILSZ=n/En Indicates the (actual or estimated) number of records after input
processing (E14, E15, INCLUDE/OMIT, SKIPREC, STOPAFT,
and Phase 1 SUM or DUPKEYS). FILSZ=n causes sort termina-
tion if n is incorrect.

Table 46. (Page 1 of 3)PARM Option Summary Chart

5–2 Syncsort MFX Programmer’s Guide

 PARM Option Name Description
FLAG FLAG and MSG control the routing of output messages.
HBSI Enables hiperbatch processing for SORTIN data set.
HBSO Places SORTOUT data set into hiperbatch.
IO Minimizes IO activity at expense of other performance measures.
See BALANCE, CPU, and ELAP. Note that these options and IO
are all mutually exclusive.
IOERR=ABE/ NOSNAP /NOIOERR Indicates how to handle I/O errors: user abend 999 with or with-
out dump, or MFX error message only.
JPn''...'' Establishes dictionary_names for substitution into MFX control
statements.
L6=n,L7=n Passes HISTOGRM data to optimize variable-length record sorts.
LIST /NOLIST LIST causes header line and control statements to be printed.
LOCALE= NONE /CURRENT /name Controls collating based on cultural environment.
MSG MSG and FLAG control the routing of messages.
MSGDD Changes the DD name of the message data set.
NOTMTOUT=RC0/RC4/RC16 Specifies the action to be taken when SORTOUT contains at least
one data record.
NULLOUT= RC0 /RC4 /RC16 Specifies the action to be taken when SORTOUT contains no
records.
OPTELAP Enables running of offloaded work on the general CP for improved
elapse time when all zIIP engines are busy.
OVFLO= RC0 /RC4 /RC16 Specifies the action to be taken if a SUM or AVG field overflows or
underflows during SUM or DUPKEYS processing.
PAD= RC0 /RC4 /RC16 Specifies the action to be taken if the non-OUTFIL SORTOUT
LRECL is larger than the output record length.
PRINT121/NOPRINT121 Changes the DCB of the message data set.
RC16=ABE/NORC16 RC16=ABE changes return code 16 to user abend 16.
RELEASE= ON /OFF Overrides the RLSE operand in the SPACE parameter of the
SORTWK DD statement(s).
RESERVE=n/nK Specifies the amount of memory reserved for the user below the
16-megabyte line.
RESERVEX=n/nK Specifies the amount of memory reserved for the user above the
16-megabyte line.
RESERVEZ=nG Specifies the amount of memory reserved for the user above the
2-gigabyte bar.
RESET/ NORESET Affects VSAM SORTOUT only.
RLSOUT/ NORLSOUT Determines whether excess space is released.

Table 46. (Page 2 of 3)PARM Option Summary Chart

Syncsort MFX Programmer’s Guide 5–3

 PARM Option Name Description
SDB= ON/ OFF/ YES/ NO/ Specifies whether system-determined blocksize should be used to
DISKONLY/ TAPEONLY/ LARGE/ select an optimum blocksize for data sets when none is provided.
SMALL/ INPUT/ LARGEONLY/
INPUTONLY
SKIPREC=n Indicates that n records should be skipped before the input file is
sorted or copied. SKIPREC is not available for PARASORT.
STOPAFT=n Sorts or copies at most n records that survive input file editing
(E15, INCLUDE/OMIT, SKIPREC etc.) STOPAFT is not available
for PARASORT.
SWCRYPT Specifies whether or not to use encryption for SORTWORK data
sets during the SORT process.
TRUNC= RC0 /RC4 /RC16 Specifies the action to be taken if the non-OUTFIL SORTOUT
LRECL is smaller than the output record length.
UNINTDS=YES/ NO Indicates if an uninitialized SORTIN or SORTINnn input file
should be processed.
VLTEST=(n/ 1 , ON /OFF /OFF4) Indicates the type of validity testing to be done when processing
variable-length records.
VLTESTI=n/ 0 Indicates action to be taken when a variable-length record does
not contain all fields referenced by INCLUDE or OMIT
processing.
VSAMEMT= NO /YES Specifies the processing of empty VSAM data sets provided as
input to a sort, merge, or copy.
ZDPRINT/ NZDPRINT Specifies whether positive summed or averaged ZD fields will be
converted to a printable format.
ZIIPPCT Enables you to define the zIIP offload percentage within a range
of 1-100.

Table 46. (Page 3 of 3)PARM Option Summary Chart

The PARM options in Table 42 are described in detail in the “MFX PARM
Options” on page 5-5.

Additional PARMs
MAXSORT

The MAXSORT feature, designed for large sorting applications, is initiated by the
MAXSORT PARM. The following additional PARMs can be specified for a MAXSORT
application: BKPTDSN, DYNATAPE, MAXWKSP, MINWKSP, NODYNATAPE,
RESTART, SORTSIZE, SORTTIME, and TAPENAME. These PARMs are described in
“Chapter 9, MAXSORT”.

5–4 Syncsort MFX Programmer’s Guide

 PARASORT

The PARASORT feature, designed to reduce elapsed time for multi-volume and/or con-
catenated tape SORTIN sort applications, is initiated by the PARASORT PARM. For
additional information on PARASORT, see “Chapter 10, PARASORT”.

DB2 Query Support

DB2 Query Support, which allows DB2 data to be passed directly into a SORT or COPY
operation without the use of setup steps or the need for user-written E15 exits, is initi-
ated by the DB2 Query Support PARM. For additional information, see “Chapter 11,
MFX DB2 Query Support”.

MULTIIN

The MULTIIN facility, which allows MFX to process multiple VSAM and non-VSAM
data sets for input, is initiated by the MULTIIN PARM. This facility is described in
“Chapter 12, Multiple Input Files”.

MFX PARM Options

BALANCE

BALANCE

Figure 276. BALANCE Format

BALANCE optimizes overall performance by balancing among CPU time, sort

elapsed time, and I/O activity to SORTIN, SORTOUT and SORTWK. If you want to
emphasize one performance measure at the possible expense of others, use CPU,
ELAP, or IO. See CPU, ELAP, and IO, below. Note that these options and
BALANCE are all mutually exclusive.

BMSG

BMSG

Figure 277. BMSG Format

BMSG enables class B messages. They will appear wherever the MSG PARM
option indicates informational messages are to be routed.

Syncsort MFX Programmer’s Guide 5–5

CENTWIN

0 
 
CENTWIN =  s 
 
f 

Figure 278. CENTWIN Format

CENTWIN defines a sliding or fixed 100-year window that determines the century
to which 2-digit year data belongs when processed by SORT, MERGE, INREC,
OUTREC or OUTFIL OUTREC control statements.
The 2-digit year data formats (Y2B, Y2C, Y2D, Y2P, Y2S and Y2Z) plus the full-
date formats (Y2T, Y2U, Y2V, Y2W, Y2X, and Y2Y) work with CENTWIN to treat a
2-digit year value as a 4-digit year. The 2-digit and full-date year data formats can
be specified on control statements as follows:
• Use SORT/MERGE control statements to correctly collate 2-digit years that
span century boundaries. For information on using the 2-digit and full-date
data formats for SORT/MERGE field specifications, see “CENTWIN Parameter
(Optional)” on page 2-71 or on page 2-239.
• Use INCLUDE/OMIT or OUTFIL INCLUDE/OMIT control statements for
correct comparisons involving 2-digit year and full-date data formats. For
information on using the 2-digit year data formats for INCLUDE/OMIT
processing, see “Specifying Field-to-Field Standard Comparisons for Year
Fields” on page 2-39. For more information on specifying full-date formats, see
pages 2-39-2-45.
• Use INREC/OUTREC or OUTFIL OUTREC control statements to convert 2-
digit year and full-date data to 4-digit printable output. For information on
using the 2-digit year data formats for OUTREC processing, see “Converting
Year Data with Century Window Processing on INREC, OUTREC, or OUTFIL
OUTREC” on page 2-162 and “Example 5” on page 2-222. For more information
on converting full-date formats, see the descriptions of the fi and fy2f,(c)
parameters on pages 2-142-2-147, Table on page 2-76, and Table 20 on page 2-
148.
In addition, two date formats, Y2ID and Y2IP, are provided for year conversion
with INREC/OUTREC and OUTFIL OUTREC. These formats work with
CENTWIN to expand a 2-digit year in packed decimal format to a 4-digit year
while maintaining the packed decimal format in the output field.
CENTWIN ensures that year data spanning centuries will be sequenced correctly.
For example, without CENTWIN processing, an ascending sort/merge would
sequence the year 01 before the year 98. With CENTWIN processing, the 01 field
could be recognized as a twenty-first century date (2001) and would thus be
sequenced after 98 (1998) for an ascending sort.

5–6 Syncsort MFX Programmer’s Guide

 The CENTWIN option generates either a sliding or fixed century window,
depending on which form of CENTWIN is used: CENTWIN=s or CENTWIN=f.
• CENTWIN=s specifies a sliding century window, which automatically advances
as the current year changes.
The variable s is a number 0 through 100. This value is subtracted from the
current year to set a century-window starting point. For example, in 1996
CENTWIN=20 would create the century window 1976 through 2075. Ten years
later in 2006, the century starting year would slide to 1986 (2006 minus 20 =
1986) and the century window would be 1986 through 2085.
The CENTWIN delivered default is s=0, which means the current year is the
starting year of a century window.
• CENTWIN=f specifies a fixed century window.
The variable f is a 4-digit year (yyyy) between 1000 and 3000. For example,
CENTWIN=1976 establishes a fixed starting year 1976 for the century window
1976 through 2075. This window will not change as the current year changes.
The century window defined by CENTWIN controls processing of year-data. If a 2-
digit year field (indicated by Y2B, Y2C, Y2D, Y2P, Y2S, Y2Z, Y2ID, Y2IP, Y2T, Y2U,
Y2V, Y2W, Y2X, or Y2Y) has a value less than the last two digits of the century
window start year, the year field will be treated as a year in the century following
the year of the century window, except for 00, which is considered to be in the same
century as the century window start year. All other 2-digit years will be treated as
in the same century as the century window start year.
For example, consider the century window 1950 through 2049. The 2-digit year
fields would be processed as follows:

Two-digit Field Processed as Year

00 2000

01 2001

49 2049

50 1950

99 1999

An ascending sort of the above sample data would produce output data in the
following sequence:

Two-digit Field Processed as Year

50 1950

99 1999

Syncsort MFX Programmer’s Guide 5–7

 00 2000

01 2001

49 2049

If CENTWIN has been specified on the SORT or MERGE control statement as well
as in the PARM field, the PARM specification has precedence.

CMP

 CPD 
CMP =  
 CLC 

Figure 279. CMP Format

CMP specifies the kind of compare operation to be used for sort/merge control fields
up to 16 bytes long, bearing the format code PD or ZD.
When CMP=CPD is used, ZD fields are PACKed and then compared. Invalid PD
data may cause a system 0C7 abend and program termination. The integrity of
fields labelled “ZD” is only guaranteed when they contain valid ZD data. Since the
zone bits (the leftmost four bits of each byte) are lost during packing, UNPKing the
field later restores only valid ZD data to its original state. Leading blanks are
transformed to leading zeros and alphabetic character data that packs to a valid
PD field is converted to valid ZD data.
CMP=CLC uses the compare logical instruction for all PD and ZD control fields. No
data validation is done and the integrity of the output is maintained.
CMP=CPD is the delivered default for the PARM. The delivered default for
VLTEST is 1, and is consistent with this default. Changing the VLTEST default
from 1 to any even number forces the use of CMP=CLC when sorting variable-
length records.
For more detailed information and sample comparisons, see the section
“Comparing PD and ZD Control Fields” on page 2-70.

5–8 Syncsort MFX Programmer’s Guide

COMMAREA

 COMMAREA(n,x) 
 
 COMMAREA(n) 
 
 COMMAREA(,x) 
 COMMAREA 
 
 NOCOMMAREA 
 

Figure 280. COMMAREA Format

COMMAREA instructs MFX to provide an area for communication between exit

programs. The size of this area is given as a decimal number n of bytes; x, a
character string at most n bytes long, designates the initial value to be stored in
this area. Regardless of the value of n, which may be between 1 and 256, x may not
exceed 89 bytes in length. (Whenever x has fewer than n characters, it will be right-
padded with blanks to a length of n.) If COMMAREA is specified via the EXEC
statement, blanks may be included within the string x. However, if COMMAREA is
specified via the $ORTPARM DD statement, intervening blanks are not allowed. In
neither case is a right parenthesis permitted since it delimits the COMMAREA
parameter.
Both n and x are optional. If either subparameter is specified, it will determine the
other: n defaults to the length of x, x defaults to n blanks. If neither x nor n is
specified, n defaults to 80 bytes, x to 80 blanks.
NOCOMMAREA is the program default: no area for communication between exit
programs is provided, although exit routines may still use the 19th word of the save
area.
Exit program access to this communication area is described in the discussion of
exit programs, see “The Exit Communication Area” on page 7-5.
COMMAREA does not apply to exits called using the 64-bit parameter list.

PARM Code Communication Area

COMMAREA(10,DEBUG) DEBUG..... (5 blanks)

COMMAREA(10) ......... (10 blanks)
COMMAREA ......... (80 blanks)
COMMAREA(,DEBUG) DEBUG
COMMAREA(80,DEBUG) DEBUG.... (75 BLANKS)

Figure 281. Examples: Coding the COMMAREA PARM

Syncsort MFX Programmer’s Guide 5–9

CORE

n 
 nK 
 
 nM 
 CORE   
  =  MAX 
 SIZE   
 MAX-n 
 MAX-nK 
 
 MAX-nM 

Figure 282. CORE Format

CORE is used to override the installation default for the amount of memory the
sort/merge is allowed to use. To specify an amount of memory, choose one entry
from each pair of braces.
Note that CORE and SIZE are synonymous. Note also that memory specification
may be a decimal number of bytes (CORE=n), a decimal multiple of K, where
K=1024 bytes (CORE=nK), or a decimal multiple of M, where M=1048576 bytes
(CORE=nM).
For simplicity, the following describes only CORE, specified in units of nK.
CORE=nK Defines a maximum memory limit of nK below the 16-
megabyte line.
CORE=MAX Assigns to the sort/merge all the available memory above
and below the 16-megabyte line.
CORE=MAX-nK Assigns to the sort/merge all the available memory above
and below the 16-megabyte line less nK bytes, which is
reserved below the 16-megabtye line.
Consult your systems staff for any installation-specific modifications to the
handling of CORE. For example, "CORE" will be limited if a maximum memory size
for MFX was set at installation time.

CPU

CPU

Figure 283. CPU Format

CPU minimizes the CPU time of each sort at the expense of sort elapsed time and I/
O activity. See BALANCE, ELAP and IO. Note that these options and CPU are all
mutually exclusive.

5–10 Syncsort MFX Programmer’s Guide

DEBUG

DEBUG

Figure 284. DEBUG Format

DEBUG produces an MFX SNAP dump in the event that a critical error forces the
sort to terminate. A SNAP dump produced in this way is of use to an MFX analyst
in debugging complex problems. See “Diagnostics and Technical Support” on
page 18-1. Note that the PSW AT ENTRY TO SNAP and general registers are
useless for debugging.

DIAG

DIAG

Figure 285. DIAG Format

DIAG turns on both the IOERR=ABE and the RC16=ABE options (see these
options for explanations).

DYNALLOC

 
d 
  (nn,mm)  
   
DYNALLOC =  (d, n ,RETRY =  OFF  [,SC=s]) 
  5, 3  
   
 
 OFF 

Figure 286. DYNALLOC Format

DYNALLOC requests the dynamic allocation of SORTWK data sets on device type
d. Specify the device type either as a decimal number (e.g., 3380) or by the system
generic name (e.g., SYSDA). Any disk device accepted for a SORTWK DD
statement can be specified. Note that if VIO is specified it will be ignored, and the
installation default for the DYNALLOC device type will be used in its place.
Note that the DYNALLOC parameter may be used alone, without any
subparameters. In this case, the DYNALLOC installation default settings are used.
For non-MAXSORT applications, n can be 1 through 255. The value n specifies the
number of SORTWK data sets that can potentially be allocated. For values of n
that are 31 or less, MFX can automatically raise the number to 32 if the application
requires. When n is 33 through 255, this value specifies the maximum number of
SORTWK data sets that can be allocated.

Syncsort MFX Programmer’s Guide 5–11

 For MAXSORT applications, n is the number of SORTWK data sets that will be
allocated. As many as 32 SORTWK data sets can be specified for MAXSORT
applications.
The delivered default for n is 3.
DYNALLOC=OFF can be specified to override a DYNALLOC=ON installation
default.
SORTWK data sets allocated by the DYNALLOC parameter normally supplement
any SORTWK data sets allocated by SORTWKnn DD statements; however, note
that there is an installation option to disable DYNALLOC if SORTWKnn DD
statements are present.
MFX uses the value specified in the RETRY parameter to request automatic
DYNALLOC retry. This facility attempts to avoid a sortwork capacity exceeded
condition when disk space is not immediately available to satisfy a DYNALLOC
request. When RETRY is specified, MFX will automatically retry a specified
number of times and wait a prescribed interval between DYNALLOC requests.
The nn in the first position designates the number of times MFX will retry a failed
DYNALLOC request. The minimum allowed is 1 and the maximum is 16. The mm
in the second position designates the number of minutes MFX waits between each
DYNALLOC request. The minimum allowed is 1 and the maximum is 15.
RETRY=OFF can be specified to override a RETRY=ON installation default.
In an environment where DFSMS manages temporary work data sets, the SC
subparameter specifies a storage class s for MFX to use when dynamically
allocating SORTWORK data sets. The storage administrator at your installation
defines the names of the storage classes you can specify. Note that an installation
written automatic class selection (ACS) routine can override the storage class you
specify. If SMS is not installed or active to manage temporary work data sets, the d
device specification will be used in the SORTWORK dynalloc request.
If DYNALLOC has been specified on the SORT control statement as well as in the
PARM field, the PARM specification will take precedence.

E15

E15=COB

Figure 287. E15 Format

E15 specifies the E15=COB option in order to include an E15 exit written in
COBOL without coding C on the MODS control statement.

5–12 Syncsort MFX Programmer’s Guide

E35

E35=COB

Figure 288. E35 Format

E35 specifies the E35=COB option in order to include an E35 exit written in
COBOL without coding C on the MODS control statement.

ELAP

ELAP

Figure 289. ELAP Format

ELAP minimizes the elapsed time (wall clock time) of each sort at the expense of
CPU time and I/O activity. See BALANCE, CPU and IO. Note that these options
and ELAP are all mutually exclusive.

EQUALS

 EQUALS 
 NOEQUALS 
 

Figure 290. EQUALS Format

EQUALS insures that the original order of equal-keyed records is preserved. These
records will be in the same order in the output file as they were in the sort input
file. For a merge, equal-keyed records from the lowest-numbered SORTINnn file
are written before those from the second input file, and so on. With NOEQUALS,
there is a random element to the order in which records with identical control fields
will appear in the output. With or without EQUALS, MERGE preserves the order
of equal-keyed records within any one data set.
If EQUALS is in effect in an application with SORTMInn data sets, the order of
equal-keyed records within each SORTMInn file will be preserved. In addition,
equal-keyed records from the lowest-numbered SORTMInn file will be written
before those from the second SORTMInn file, and so on.
When used in conjunction with SUM or DUPKEYS, EQUALS indicates which of
the equal-keyed records will be preserved, containing the sum or DUPKEYS
minimum, maximum, or average values: the record occurring first in SORTIN (for
a sort), or drawn from the SORTINnn data set with the lowest nn number (for a
merge) will contain the calculated fields.

Syncsort MFX Programmer’s Guide 5–13

 The EQUALS option can also be specified on the SORT/MERGE control statement.
The specification on the control statement takes precedence over the specification
in the PARM field.

EXTCOUNT

EXTCOUNT

Figure 291. EXTCOUNT Format

EXTCOUNT enables special processing to accommodate applications that have

record counts that exceed MFX’s default internal limit.
By default, the internal limit on the number of records that can be sorted for
variable-length data or for a sort application that uses the EQUALS option is
4,294,967,295 records. Specifying EXTCOUNT increases the internal limit to
140,737,488,355,327 records. Fixed-length sorts without EQUALS, and all merges
and copies, have automatic support for the maximum number of records allowed by
the EXTCOUNT parameter.
Note that additional SORTWK space may be required when specifying the
EXTCOUNT parameter with a VL sort or a fixed-length sort with EQUALS. The
additional SORTWK space is 2 bytes per record. This amount can add a significant
percentage to the SORTWK space needs if the LRECL of the records is small.
(Small LRECLs are typical of files with an extremely large number of records).
Therefore, when using EXTCOUNT with a VL sort or a fixed-length sort with
EQUALS, insure that the extra SORTWK space will be available.
Performance will usually be improved if the EXTCOUNT option is not in effect.
Therefore, EXTCOUNT should be used only when appropriate to the application.
If the record limit is exceeded, MFX will issue a critical error message and
terminate the application.

FILSZ

n 
FILSZ =  
 En 

Figure 292. FILSZ Format

FILSZ indicates the actual (FILSZ=n) or estimated (FILSZ=En) decimal number of

records to be sorted, taking into account all record additions and deletions due to
an E14 or E15 exit routine, the INCLUDE/OMIT control statement, the SUM or
DUPKEYS control statement and the SKIPREC and STOPAFT parameters.
FILSZ=n instructs MFX to terminate with an error message unless exactly n
records are to be sorted. Since the number of records deleted by SUM or DUPKEYS

5–14 Syncsort MFX Programmer’s Guide

 processing in Phase 1 is indeterminate and may not be reproducible, much less
predictable, only the estimated En value should be used if a SUM or DUPKEYS
control statement is present.
The FILSZ option can also be specified on the SORT control statement. The
specification in the PARM field will take precedence over that on the control
statement.

FLAG

 FLAG(I) 
 
 FLAG(U) 
 
 NOFLAG 

Figure 293. FLAG Format

FLAG controls the routing of output messages. The MSG option, which handles
messages more comprehensively, is explained later in this section. The format of
the FLAG option is given below.
Specify FLAG(I) to route all messages to the data set specified by the SYSOUT DD
statement and only critical messages to the console. (This is the same as the
MSG=SC PARM.)
Specify FLAG(U) to route critical messages only to both the data set specified by
the SYSOUT DD statement and the console. (This is the same as the MSG=CB
PARM.)
Specify NOFLAG to route critical messages only to the console, no messages to the
data set specified by the SYSOUT DD statement. (This is the same as the MSG=CC
PARM.)

HBSI

HBSI

Figure 294. HBSI Format

HBSI turns on hiperbatch processing for SORTIN data sets. To benefit from
hiperbatch processing, the SORTIN data set should already reside in hiperbatch.
Although hiperbatch does provide significant improvements in elapsed time, it
causes some degree of degradation in other system resources. If you use HBSI and
the SORTIN data set does not reside in hiperbatch, you may experience some
system degradation while not realizing any of the benefits that accompany
hiperbatch processing.

Syncsort MFX Programmer’s Guide 5–15

HBSO

HBSO

Figure 295. HBSO Format

HBSO turns on hiperbatch processing for SORTOUT data sets. HBSO benefits only
subsequent job steps that utilize hiperbatch to access this data set.

IO

IO

Figure 296. IO Format

IO minimizes the I/O activity of each sort at the expense of sort elapsed time and
CPU time. See BALANCE, CPU and ELAP. Note that these options and IO are all
mutually exclusive.

IOERR

 IOERR=ABE 
 
 NOIOERR 
 
 IOERR=NOSNAP 

Figure 297. IOERR Format

IOERR specifies IOERR=ABE to receive user abend 999 if an I/O error should
occur. This abend will cause the job step to terminate, producing a diagnostic dump.
NOIOERR is the program default. If this option is in effect, MFX will, in the event
of an I/O error, terminate with either a return code of 16 or a user abend 16,
depending on the RC16 option that is used.
Specify IOERR=NOSNAP to receive a user abend 999 if an I/O error should occur.
This abend will cause the step to terminate, but no diagnostic dump will be
produced, unless the DEBUG option is in effect.

JPn

JPn''...'' where n is from 0 to 9

Figure 298. JPn Format

JPn is used to establish dictionary_names for symbolic substitution in MFX control

statements. JCL SET and PROC symbols, system symbols and text strings may all

5–16 Syncsort MFX Programmer’s Guide

 be used in the JPn PARM to establish up to 10 character string dictionary_names.
See “Using JCL SET and PROC Symbols to Create Dictionary_Names” on page 13-
29 for a complete description.

L6

L6=n

Figure 299. L6 Format

L6 indicates the average number of bytes of work space each record will need,
overriding (if present) the l6 parameter of the RECORD control statement. The
decimal value n of the optional L6 parameter is provided by the HISTOGRM utility
program. If neither L6 nor l6 is provided, MFX will estimate this value.
L6 is only used for sorting variable-length records. It is ignored by merge, and copy
applications.

L7

L7=n

Figure 300. L7 Format

L7 indicates the segment length that MFX should use for maximum sorting
efficiency. The decimal value n of the optional L7 parameter is provided by the
HISTOGRM utility program. (A segment is a fixed-length area used to contain all
or part of a variable-length record.) The L7 value overrides (if present) the l7
parameter of the RECORD control statement. If neither L7 nor l7 is provided, MFX
will estimate this value.
L7 is only used for sorting variable-length records. It is ignored by merge, and copy
applications.

LIST

 LIST 
 
 NOLIST 

Figure 301. LIST Format

LIST, the default for the sort/merge program, causes header lines and control
statements to be listed with the SYSOUT data set (in all likelihood, at the printer)
for both JCL- and program-initiated executions. If NOLIST is specified, the control
statements and header lines will not appear with this data set.

Syncsort MFX Programmer’s Guide 5–17

LOCALE

 NONE 
 
LOCALE =  CURRENT 
 
 name 

Figure 302. LOCALE Format

LOCALE controls cultural environment processing, allowing you to choose an

alternative set of collating rules based on a specified national language. For SORT/
MERGE processing, the alternative collating applies to character (CH) fields. For
INCLUDE/OMIT comparison processing, the alternative collating applies to
character fields and hexadecimal constants compared to character fields.
MFX employs the callable services of IBM’s Language Environment for z/OS to
collate data in a way that conforms to the language and conventions of a selected
locale. A locale defines single and multi-character collating rules for a cultural
environment. Numerous predefined locales are available.
NONE, the default setting for LOCALE, results in normal EBCDIC collating.
CURRENT directs MFX to use the locale active when MFX begins.
name is the name of a supplied or user-defined locale that is to be active during
MFX processing. A locale name may be up to 32 characters and is not case
sensitive. The locale active just before MFX processing begins will be restored when
MFX processing completes. The following is a list of locales provided with the IBM
National Language Resources Feature of LE/370.

5–18 Syncsort MFX Programmer’s Guide

 Locale
Language Country
Name

DA_DK Danish Denmark

DE_CH German Switzerland

DE_DE German Germany

EL_GR Greek Greece

EN_GB English United Kingdom

EN_JP English Japan

EN_US English United States

ES_ES Spanish Spain

FI_FI Finnish Finland

FR_BE French Belgium

FR_CA French Canada

FR_CH French Switzerland

FR_FR French France

IS_IS Icelandic Iceland

IT_IT Italian Italy

JA_JP Japanese Japan

NL_BE Dutch Belgium

NL_NL Dutch Netherlands

NO_NO Norwegian Norway

PT_PT Portuguese Portugal

SV_SE Swedish Sweden

TR_TR Turkish Turkey

Table 47. Defined Locales

Notes:

1. Make sure the JCL gives MFX access to the library that contains the loadable locale
routines. For the supplied locales, these are the dynamically loadable routines in the

Syncsort MFX Programmer’s Guide 5–19

 IBM AD/Cycle LE/370 library. For more information, see the IBM publication Language
Environment for z/OS & VM Installation and Customization Guide, SC26-4817.

2. If locale processing is used for fields specified in a SORT or MERGE control statement,
VLTEST=1 will be forced on in addition to any other VLTEST options in effect.
VLTEST=1 will cause MFX to terminate if a variable-length input record does not
contain all SORT/MERGE control fields.

3. Although locale processing can improve performance compared to external collating

routines, it should be used only when necessary. Locale processing can significantly
degrade SORT/MERGE and INCLUDE/OMIT performance compared to normal
collating.

4. An E61 exit cannot be used with locale processing.

5. Locale processing requires additional main storage to support the use of the IBM
Language Environment facilities. For those jobs that use locale, the below-the-line
region size should be increased by 1000K to accommodate the storage needs of the
Language Environment modules.

MSG

 AB 
 AC 
 
 AP 
 
 CB 
 
 CC 
MSG =  
 CP 
 
 NO 
 PC 
 
 SC 
 
 SP 

Figure 303. MSG Format

MSG indicates where MFX messages are to be routed. The MSG codes assume that
the printer is specified for the message data set; if a device other than the printer is
specified for this data set, messages described as routed to the printer will be
routed to this other device instead.
AB causes all messages to be routed both to the printer and to the console.
AC causes all messages to be routed to the console, none to the printer.
AP causes all messages to be routed to the printer, none to the console. This is
the program default.

5–20 Syncsort MFX Programmer’s Guide

 CB causes only critical messages to be routed to the printer and to the console.
(This is the same as the FLAG(U) option.)
CC causes only critical messages to be routed to the console, no messages to the
printer. (This is the same as the NOFLAG option.)
CP causes only critical messages to be routed to the printer, no messages to the
console.
NO causes no messages to be routed to either the printer or the console.
PC causes all messages to be routed to the printer and to the console.
SC causes only critical messages to be routed to the console, all messages to the
printer. (This is the same as the FLAG(I) option.)
SP causes only critical messages to be routed to the printer, all messages to the
console.

MSGDD

 SYSOUT 
MSGDD =  
 xxxxxxxx 

Figure 304. MSGDD Format

The program default for the DD name of the message data set is SYSOUT. To
assign a different DD name, substitute any valid DD name for xxxxxxxx.

NOTMTOUT

 RC0 
 
NOTMTOUT=  RC4 
 
 RC16 

Figure 305. NOTMTOUT Format

NOTMTOUT specifies the action to be taken when SORTOUT in a sort, merge, or

copy application contains at least one data record.
RC0 The delivered default instructs MFX to issue a return code of 0 if not over-
ridden by a higher return code set for another reason.
RC4 Instructs MFX to issue a WER495I warning message and continue process-
ing. A return code of 4 will be issued if not overridden by a higher return
code set for another reason.
RC16 Instructs MFX to issue a WER495A message and terminate processing with
a return code of 16.

Syncsort MFX Programmer’s Guide 5–21

 NOTMTOUT will be ignored for a BetterGener application.

NULLOUT

 RC0 
 
NULLOUT =  RC4 
 
 RC16 

Figure 306. NULLOUT Format

NULLOUT specifies the action to be taken when SORTOUT in a sort, merge, or

copy application contains no data records.
RC0 The delivered default instructs MFX to issue a return code of 0 if not over-
ridden by a higher return code set for another reason.
RC4 Instructs MFX to issue a WER461I warning message and continue process-
ing. A return code of 4 will be issued if not overridden by a higher return
code set for another reason.
RC16 Instructs MFX to issue a WER461A message and terminate processing with
a return code of 16.
NULLOUT will be ignored for a BetterGener application.

OPTELAP

 YES 
OPTELAP =  NO 
 

Figure 307. OPTELAP Format

OPTELAP enables you to run offloaded work on the general CP to improve the
elapse time when all zIIP engines are busy. When OPTELAP=YES, then MFX will
not attempt zIIP offloads when all zIIP engines are busy.
The delivered default is NO.

OVFLO

 RC0 
 
OVFLO =  RC4 
 
 RC16 

Figure 308. OVFLO Format

5–22 Syncsort MFX Programmer’s Guide

 OVFLO specifies the action to be taken if a summed or averaged field overflows or
underflows during SUM, DUPKEYS SUM, or DUPKEYS AVG processing.
RC0 The delivered default instructs MFX to issue a WER049I warning message
and continue processing. A return code of 0 will be returned if not overrid-
den by a higher return code set for another reason. The WER049I will only
be issued on the first occurrence of the overflow or underflow.
RC4 Instructs MFX to issue a WER049I warning message and continue process-
ing. A return code of 4 will be issued if not overridden by a higher return
code set for another reason. The WER049I will only be issued on the first
occurrence of the overflow or underflow.
RC16 Instructs MFX to issue a WER049A message and terminate processing with
a return code of 16.

PAD

 RC0 
 
PAD =  RC4 
 
 RC16 

Figure 309. PAD Format

PAD specifies the action to be taken if the LRECL defined in the JCL for a non-
OUTFIL SORTOUT is larger than the SORTIN/SORTINnn LRECL or the
internally processed record length when the SORTIN/SORTINnn LRECL is
modified by features.
RC0 The delivered default instructs MFX to issue a WER462I message, pad
fixed-length output records with binary zeros, and issue a return code of
zero.
RC4 Instructs MFX to issue a WER462I message and pad fixed-length output
records with binary zeros. A return code of 4 will be issued if not overridden
by a higher return code set for another reason.
RC16 Instructs MFX to issue a WER462A message and terminate processing with
a return code of 16.
Note that for a BetterGener application PAD will be ignored. The installation
parameter SOPADGN will control processing for these applications.
PAD will be ignored in applications in which the SORTIN/SORTINnn or
SORTOUT is a VSAM data set.

Syncsort MFX Programmer’s Guide 5–23

PRINT121

 PRINT121 
 
 NOPRINT121 

Figure 310. PRINT121 Format

PRINT121 changes MFX’s DCB default for the message data set to the following:
DCB=(LRECL=121,BLKSIZE=121,RECFM=FA)
This PARM is useful when the application includes a COBOL exit which uses
DISPLAY, EXHIBIT, or TRACE instructions. (These macros will otherwise cause
conflicts between program and sort/merge messages.) An alternative is provided by
the MSGDD parameter, used to change the name of the MFX message data set.
The MFX program default for the message file’s DCB is:
DCB=(LRECL=125,BLKSIZE=882,RECFM=VBA)
If PRINT121 is set as an installation default, NOPRINT121 can be used as an
override.
PRINT121 is automatically implemented for all program-initiated sort/merges.

RC16

 RC16=ABE 
 NORC16 
 

Figure 311. RC16 Format

RC16=ABE will cause the sort to issue user ABEND 16 instead of return code 16.
The sort step is abnormally terminated without a dump and subsequent job steps
are generally flushed by the operating system. However, JCL may specify job
step(s) to be executed only in the event of an ABEND.
The delivered default is NORC16; unsuccessful completion of the sort causes a
return code of 16 to be passed to the operating system or the invoking program.

RELEASE

 ON 
RELEASE =  
 OFF 

Figure 312. RELEASE Format

5–24 Syncsort MFX Programmer’s Guide

 RELEASE=ON (the program default) turns on the RLSE operand in the SPACE
parameter of the SORTWK DD statements. This will cause unused space to be
released from sortwork units during execution time.
Specify RELEASE=OFF to instruct MFX to turn off the RLSE operand in the
SPACE parameter of the SORTWK DD statement. In this case, MFX will not
release unused space from the sortwork units during sort execution.

RESERVE

n 
 
RESERVE =  nK 
 
 nM 

Figure 313. RESERVE Format

RESERVE sets aside a specified amount of memory below the 16-megabyte line for
the user. This parameter takes effect only when CORE=MAX is in effect.
The memory may be specified as a decimal number of bytes (RESERVE=n), a
decimal multiple of K, where K=1024 bytes (RESERVE=nK), or a decimal multiple
of M, where M=1048576 bytes (RESERVE=nM).

RESERVEX

n 
 
RESERVEX =  nK 
 
 nM 

Figure 314. RESERVEX Format

RESERVEX reserves a specified amount of memory above the 16-megabyte line for
the user. This parameter takes effect only when CORE=MAX is in effect.
The memory may be specified as a decimal number of bytes (RESERVEX=n), a
decimal multiple of K, where K=1024 bytes (RESERVEX=nK), or a decimal
multiple of M, where M=1048576 bytes (RESERVEX=nM).

Syncsort MFX Programmer’s Guide 5–25

RESERVEZ

RESERVEZ = { nG }

Figure 315. RESERVEZ Format

RESERVEZ reserves a specified amount of memory above the 2-gigabyte bar for
the user.
The memory can be specified as a decimal number of G for a gigabyte Memory
Object, where G=1073741824 bytes (RESERVEZ=nG).
For example: 0G indicates no reserved Memory Object; 255G indicates reserving a
255 gigabyte Memory Object.
Valid values are 0-255.

RESET

 RESET 
 NORESET 
 

Figure 316. RESET Format

RESET will prevent VSAM from treating output data sets as MOD data sets, if an
output VSAM file was created using the REUSE option.

RLSOUT

 RLSOUT 
 NORLSOUT 
 

Figure 317. RLSOUT Format

RLSOUT releases all excess primary and secondary space from each output DASD
file when the parm is specified and DISP=NEW is specified on output data set
statements.
NORLSOUT is the program default. In this case, MFX does not release any excess
space on these output files.

5–26 Syncsort MFX Programmer’s Guide

SDB

 ON 
 
 OFF 
 YES 
 
 NO 
 
 DISKONLY 
 
SDB =  TAPEONLY 
 LARGE 
 
 SMALL 
 
 INPUT 
 
 LARGEONLY 
 
 INPUTONLY 

Figure 318. SDB Format

SDB specifies whether system-determined blocksize should be used to select an

optimal blocksize for output data sets when none is provided. This parameter will
automatically provide a blocksize that will most efficiently utilize the space on the
output device.
SDB=ON or YES enables the use of system-determined blocksize for both tape and
new or previously allocated but unopened DASD output data sets except in the
following conditions:
• A blocksize is found in the JCL DCB BLKSIZE specification or, in the case of a
DISP=MOD tape data set, it is derived from an available tape label.
• The output file is a VSAM data set.
If the output data set is on DASD, the blocksize selected will be based upon the
RECFM and LRECL, either specifically provided or determined from the usual
analysis of SORTIN or RECORD statement attributes. For example, the blocksize
selected for a blocked output data set assigned to a 3380 or 3390 DASD device will
represent a size as close to half-track blocking as possible.
If the output file is a tape data set, the blocksize will be determined from the
RECFM and LRECL in conjunction with the following rules:
• RECFM of F or FS: BLKSIZE=LRECL
• RECFM of FB or FBS and LABEL type is not AL: BLKSIZE=highest multiple
of LRECL that is less than or equal to 32760.
• RECFM of FB and LABEL type is AL: BLKSIZE=highest multiple of LRECL
that is less than or equal to 2048.
• RECFM of V, VS, D: BLKSIZE=LRECL +4
• RECFM of VB, VBS: BLKSIZE=32760

Syncsort MFX Programmer’s Guide 5–27

 • RECFM of DB: BLKSIZE=2048
If SDB=OFF or NO is specified, MFX will not use system-determined blocksize.
The blocksize, if unavailable, will be determined from SORTIN if the SORTIN and
output data set LRECLs are the same, otherwise MFX will select an appropriate
blocksize.
If SDB=DISKONLY is specified, MFX will use system-determined blocksize only
for disk output data sets.
If SDB=TAPEONLY is specified, MFX will use system-determined blocksize only
for tape output data sets.
SDB=LARGE enables the use of system-determined blocksize for both tape and
DASD output data sets, as with SDB=ON. Additionally, SDB=LARGE enables
selection of a system-determined blocksize greater than 32760 for eligible tape
output data sets if not restricted by the system BLKSZLIM value.
SDB=SMALL has the same meaning as SDB=ON.
SDB=INPUT enables the use of system-determined blocksize for both tape and
DASD output data sets, as with SDB=ON. Additionally, if an input tape data set
has a blocksize greater than 32760, SDB=INPUT enables selection of a system-
determined blocksize greater than 32760 for eligible tape output data sets if not
restricted by the system BLKSZLIM value. SDB=INPUT is the default.
SDB=LARGEONLY enables the use of system-determined blocksize for tape output
data sets only, as with SDB=TAPEONLY. Additionally, SDB=LARGEONLY enables
selection of a system-determined blocksize greater than 32760 for eligible tape
output data sets if not restricted by the system BLKSZLIM value.
SDB=INPUTONLY enables the use of system-determined blocksize for tape output
data sets only, as with SDB=TAPEONLY. Additionally, if an input tape data set has
a blocksize greater than 32760, SDB=INPUTONLY enables selection of a system-
determined blocksize greater than 32760 for eligible tape output data sets if not
restricted by the system BLKSZLIM value.

SKIPREC

SKIPREC=n

Figure 319. SKIPREC Format

SKIPREC=n instructs the sort to skip a decimal number n of records before sorting/
copying the input file. The records skipped are deleted from the input file before
E15 and INCLUDE/OMIT processing is begun.
If SKIPREC=n has been specified on the SORT/MERGE control statement as well
as in the PARM field, the PARM specification will take precedence.
SKIPREC is not compatible with a merge application unless using FIELDS=COPY.

5–28 Syncsort MFX Programmer’s Guide

STOPAFT

STOPAFT=n

Figure 320. STOPAFT Format

The STOPAFT=n parameter specifies the number of records to be sorted or copied.

These will be the first n records after E15, INCLUDE/OMIT and SKIPREC
processing, if specified, have completed.
If STOPAFT=n has been specified on the SORT/MERGE control statement as well
as in the PARM field, the PARM specification will take precedence.
STOPAFT is not compatible with merge applications, unless using FIELDS=COPY.

SWCRYPT

 YES 
SWCRYPT =  NO 
 

Figure 321. SWCRYPT Format

SWCRYPT specifies whether or not to use encryption for SORTWORK data sets
during the SORT process.
SWCRYPT=YES enables the use of encryption for SORTWORK data sets.
SORTWORK data sets will only be encrypted when SORTIN or SORTOUT are
encrypted, or when neither are present. Otherwise, SORTWORK encryption is
disabled.
SWCRYPT=NO disables the use of encryption for SORTWORK data sets.
The delivered default is NO.

TRUNC

 RC0 
 
TRUNC =  RC4 
 
 RC16 

Figure 322. TRUNC Format

TRUNC specifies the action to be taken if the LRECL defined in the JCL for a non-
OUTFIL SORTOUT is smaller than the SORTIN/SORTINnn LRECL or the
internally processed record length when the SORTIN/SORTINnn LRECL is
modified by features.

Syncsort MFX Programmer’s Guide 5–29

 RC0 The delivered default instructs MFX to issue a WER462I message, truncate
the output records, and issue a return code of zero.
RC4 Instructs MFX to issue a WER462I message and truncate the output
records. A return code of 4 will be issued if not overridden by a higher return
code set for another reason.
RC16 Instructs MFX to issue a WER462A message and terminate processing with
a return code of 16.
Note that for a BetterGener application TRUNC will be ignored. The installation
parameter SOTRNGN will control processing for these applications. TRUNC will
be ignored in applications in which the SORTIN/SORTINnn or SORTOUT is a
VSAM data set.

UNINTDS

 YES 
UNINTDS =  
 NO 

Figure 323. UNINTDS Format

UNINTDS indicates how MFX should process a non-VSAM uninitialized DASD

input data set in a non-SMS environment. An uninitialized data set is one that has
been created but never successfully opened and closed for output. In an SMS
environment, uninitialized data sets are always processed as valid empty files.
UNINTDS=YES indicates that an uninitialized data set should be processed as an
empty file. If an uninitialized multi-volume data set has the DS1IND80 and
DS1IND02 flags off in the format-1 DSCB of the first volume and the number of
data extents is non-zero, MFX will open the data set for output to set an end-of-file
mark before the data set is used for input.
UNINTDS=NO indicates that MFX should terminate with a WER400A critical
message if an uninitialized data set is provided as input.

VLTEST

 ,ON 
n   
VLTEST = (    ,OFF  )
1
   
 ,OFF4 

Figure 324. VLTEST Format

VLTEST allows you to do the following when variable-length records are processed:
• Choose the type of record length validity testing to be performed.

5–30 Syncsort MFX Programmer’s Guide

 • Choose whether or not to verify the correct sequence of segments in variable-
length spanned records.
Record length validity testing may be performed in all types of applications: sort,
merge, copy, and BetterGener. Segment sequence checking may only be done during
sort and merge applications.
The first subparameter of the VLTEST PARM is a number n that instructs MFX in
the type of validity testing to be performed on variable-length records. Choosing a
validity test instructs MFX to terminate with a critical error (outlined in
WER027A, WER160A or WER167A) in the event of an illegal condition.
A primary use of VLTEST instructs the sort/merge in the handling of “short”
variable-length records, i.e., records not long enough to contain all of the control
fields specified in the SORT/MERGE control statement. The delivered default for
VLTEST is 1.
When VLTEST is set to an even number, MFX will accept short variable-length
records, padding them with binary zeros to the length of the sort key for the sort
compare process. In order to prevent system 0C7 abends due to the binary zero
padding, the CMP PARM is automatically set to CMP=CLC in these cases. The
binary zeros are removed from the record, restoring it to its original state, as the
output record is being written.

0* No record length validity testing of variable-length records.

1 If any input record does not contain all SORT/MERGE control fields, ter-
minate. This is the default.
2* If any input record is longer than the maximum LRECL or l2 value, termi-
nate.
3 If either or both of the conditions in tests 1 and 2 are satisfied, terminate.
4* If any input record is longer than the output LRECL or l3 value, termi-
nate.
5 If either or both of the conditions in test 1 or 4 are satisfied, terminate.
6* If any input record is longer than the maximum input LRECL or l2 value,
or longer than the output LRECL or both, terminate.
7 If any of the conditions in test 1, 2, or 4 are satisfied, terminate.
* These values force the use of CMP=CLC for variable-length input.

Table 48. Values of n for VLTEST Option

The second subparameter allows you to specify whether or not MFX should verify
that the sequence of segments is correct in each variable-length spanned record
during sort and merge applications. ON is the delivered default and signals that
the segment sequence should be verified. If OFF is selected, all illogical record
segments encountered in the input file will be eliminated and message WER464I

Syncsort MFX Programmer’s Guide 5–31

 will be produced. If OFF4 is selected, the processing described for OFF will occur,
but in addition if an illogical segment is found, a return code of 4 will be returned if
not overridden by a higher return code set for another reason.
The second subparameter does not apply during copy applications. In a copy
application all illogical record segments encountered in the input file will be
eliminated.
Note: If an illegal condition is detected during a validity test and segment
sequence checking is on, message WER182A will be issued.

VLTESTI

0 
 
VLTESTI =  1 
 
2 

Figure 325. VLTESTI Format

VLTESTI specifies to MFX how to process variable-length records that do not

contain all specified INCLUDE or OMIT fields. VLTESTI applies to the INCLUDE
and OMIT control statements as well as OUTFIL and JOINKEYS INCLUDE/
OMIT processing. It does not apply to the WHEN and KEYBEGIN subparameters
of the IFTHEN parameter of the INREC, OUTREC, and OUTFIL control
statements. It also does not apply to the TRLID subparameter of the IFTRAIL
parameter of the OUTFIL control statement.
The delivered default of 0 instructs MFX to terminate if a record does not
completely contain all INCLUDE or OMIT fields. A WER250A critical error
message is generated to indicate this condition.
When VLTESTI=1 is specified, a record that does not completely contain all
INCLUDE/OMIT fields is treated as having failed the comparison. MFX will omit
the record if INCLUDE is being used or include the record if OMIT has been
specified.
When VLTESTI=2 is specified, MFX will treat comparisons to fields not completely
contained within the record as false and decide a record’s status for inclusion or
omission from fields that are available. If all fields are not present, the record will
be processed as having failed the comparison. MFX will omit the record if
INCLUDE is being used or include the record if OMIT has been specified.

5–32 Syncsort MFX Programmer’s Guide

VSAMEMT

 NO 
VSAMEMT =  
 YES 

Figure 326. VSAMEMT Format

VSAMEMT specifies the processing of empty VSAM input data sets.

If you specify VSAMEMT=YES, an empty VSAM data set will be processed as a
legitimate data set containing no records.
The delivered default, VSAMEMT=NO, instructs MFX to terminate with a
WER254A critical error if an empty VSAM data set is specified for input.

ZDPRINT

 ZDPRINT 
 
 NZDPRINT 

Figure 327. ZDPRINT Format

The ZDPRINT option applies to the SUM, DUPKEYS SUM, and DUPKEYS AVG
features.
ZDPRINT specifies if positive ZD summed or averaged results are to be converted
to printable numbers. ZDPRINT, the default, enables conversion to printable
format. NZDPRINT prevents the conversion.
This option determines whether the sign byte of a positive summed or averaged ZD
field will be converted to a printable format. More precisely, the option specifies
whether the zone of the last digit should be changed from a hexadecimal C to a
hexadecimal F.

ZIIPPCT

 nn 
ZIIPPCT =  100 
 

Figure 328. ZIIPPCT Format

ZIIPPCT option enables you to define the zIIP offload percentage. The zIIP
percentage valid value nn is in the range of 1-100. The default percent is 100.

Syncsort MFX Programmer’s Guide 5–33

5–34 Syncsort MFX Programmer’s Guide
Chapter 6 Invoking MFX from a
Program

Programming Flexibility vs. Performance

The sort/merge can be invoked by an executing program written in COBOL, PL/1 or
assembler language. However, since most invoked sorts utilize inline exits
(typically through the COBOL SORT verb) and so are handicapped by the calling
program’s I/O techniques and memory allocations, sort performance may be
degraded by this mode of initiation. Whenever performance is an important
consideration, the sort should be initiated through the EXEC job control statement.
Additional programming flexibility is provided by exits which can be separately
compiled and link-edited. These may be coded in COBOL, FORTRAN, REXX, and
assembler language. Exits may also be written in PL/1 provided that MFX is
invoked by a PL/1 program.

Syncsort MFX Programmer’s Guide 6–1

DD Statements
The DD statements included in the Table of DD Statements for Invoked Sort/Merge
are those which may be required when invoking the sort.
DD Statement Usage When not required
//$ORTPARM DD Used to override or add to control If no control statements or PARMs
statements or PARMs supplied from are being overridden or added.
an invoking program.
//SORTIN DD SORT input data set. If there is an E32 exit routine or if
running a merge or join application.
//SORTINnn DD MERGE input data sets. If there is an E32 exit routine of if
//SORTINn DD running a sort or copy.
//SORTJNF1 DD JOINKEYS input data sets If the JOINKEYS feature is not in
//SORTJNF2 DD use.
//SORTWKxx DD Work area definition. Defines the For incore sort, MERGE, COPY, or
//SORTWKn DD intermediate storage for a sort. when using DYNALLOC.
//SORTOUT DD Output data set. If there is an E35 exit routine.
or
//SORTOFxx DD
or When no OUTFIL control statement
//SORTOFx DD Alternate output data set(s). is present.

//SORTXDUP DD Output data set for records deleted When the XDUP parameter of the
by DUPKEYS. DUPKEYS control statement is not
used.
//SORTXSUM DD Output data set for records deleted When the XSUM parameter of the
by SUM. SUM control statement is not used.
//SYSOUT DD Message data set. When all messages are routed to con-
sole or are disabled.

Table 49. DD Statements for Invoked Sort/Merge

Invoking the Sort/Merge from an Assembler Program

Assembler invocation is accomplished by means of the ATTACH, LINK, or XCTL
macro instruction. MFX control statements are coded as character-string operands
of Assembler DC operations. The calling program passes to the sort/merge a
pointer containing the address of either a 24-bit or extended 31-bit or 64-bit
parameter lists. This list contains the addresses of the control statement images to
be used by the sort. It may also contain other information such as the addresses of
E15, E32, and E35 exit routines. Note that when using any of the parameter lists,
the control statement images can be passed from $ORTPARM.

6–2 Syncsort MFX Programmer’s Guide

Macro Instructions
The choice of macro determines the linkage relationship between the calling
program and the sort/merge load module. The linkage relationship established by
ATTACH precludes the use of the Checkpoint-Restart feature; do not code CHKPT/
CKPT on the SORT/MERGE control statement when invoking the sort/merge with
the ATTACH macro. With XCTL, care must be taken to ensure that the storage
area for the parameter list and other sort control information does not reside in the
module issuing the macro - XCTL will delete this module from memory. This
problem can be circumvented in two ways. Either (1) place the parameter list and
additional control information in the task that attaches the module issuing the
XCTL; or (2) have the module issuing the XCTL macro issue a GETMAIN macro
instruction first, and place all of the sort/merge control information in the main
storage area it obtains. None of the above restrictions apply when using the LINK
macro.
The sort/merge DD statements are placed with the JCL of the job step that issues
the macro. The EP parameter is specified as SORT whether MFX is to be used for
sorting, merging, or copying. With ATTACH, the ECB or EXTR is usually required.

Coding the Sort/Merge Control Statements

MFX control statements are introduced by the invoking programs as character
operands in DC operations. Although it is generally true that all control statements
supported for a JCL-initiated Disk Sort/MAXSORT are available to invoked
applications, these exceptions should be noted:
• A MODS control statement cannot be used for an E32 exit and is not required
for an E15 or E35 exit. (An E32, E15 and/or E35 exit routine may be coded in
line with the invoking program with their addresses passed to the sort in the
parameter list. If a 31-bit or 64-bit parameter list is being used, an E18 and/or
E39 exit routine address may also be passed.)
• A RECORD control statement is required if an inline E15 exit routine address
is provided. Its LENGTH parameter is required whenever an inline E15 or E35
exit routine is used. For a full description of record length parameters, see
“RECORD Control Statement” on page 2-227.
• The END control statement is not used.
The actual coding of the control statements is the same, except that:
• No comments or labels are permitted within the DC operand.
• Continuation characters are not called for, since the statements are not in card-
image format.
For the 24-bit parameter list, each control statement DC instruction should be
labeled and followed by a DC C' ' instruction so that the beginning and ending
addresses of the control statement can be referenced in the parameter list.

Syncsort MFX Programmer’s Guide 6–3

 SORTBEG DC C' SORT FIELDS=(31,5,CH,A)'
SORTEND DC C' '
RECDBEG DC C' RECORD TYPE=F,LENGTH=(80,,120)'
RECDEND DC C' '

Figure 329. Sample Control Statement Images for the 24-bit Parameter List

Note that the 24-bit invoking parameter list and pointer word, as well as the
control statement images, must be below the 16 megabyte line.
As in a JCL-initiated sort/merge, control statements must begin with at least one
blank.
For the 31-bit and 64-bit invoking parameter lists, the control statement images
will be pointed to by the first word of the parameter list and are organized like the
control statement images for a 24-bit parameter list. Note, however, that only the
first DC instruction requires a label since only the start and end of the list need be
referred to. The control statements must be separated by one or more blanks; that
is, each control statement must be followed by a blank. A blank before the first
statement is optional; however, a blank after each statement is required. Labels,
comment statements, and comment fields must not be coded. Each control
statement, except the last, must have at least one parameter.

CARDLEN DC 0H
DC Y(CARDEND-CARDBEG)
CARDBEG DC C'SORT FIELDS=(31,5,CH,A)'
DC C' '
DC C'RECORD TYPE=F,LENGTH=(80,,120)'
DC C' '
CARDEND EQU *

Figure 330. Sample Control Statement Images for the 31-bit Parameter List

The 24-Bit Parameter List

The parameter list is used to pass information from the calling program (e.g., the
addresses of the DC control statement images) to the sort/merge.
In order to pass the parameter list, it is necessary to load the address of a fullword
pointer into Register 1. Code X'80' in the pointer’s first byte and the address of the
parameter list’s byte count in its last three bytes. The byte count must be located in
the last 2 bytes of the first fullword entry in the parameter list. It contains the
hexadecimal number of bytes remaining in the list -- do not include these 2 bytes in
the count.

6–4 Syncsort MFX Programmer’s Guide

 The first seven words of the parameter list are required and must be coded in the
exact order shown. Note that whenever the address of an exit routine is supplied in
the sixth or seventh word of the parameter list, that exit may not be specified in the
MODS control statement image. Parameter list entries following the seventh
fullword are optional and can be specified in any order. All values not specified in
the chart as EBCDIC-coded are to be given in hexadecimal notation. For the most
part, these are addresses; the ending address refers to the blank coded immediately
after the control statement in question, as indicated in the Sample Invoked Sort.

Syncsort MFX Programmer’s Guide 6–5

 REGISTER 1 POINTER (Fullword)
ADDRESS OF POINTER X'80' Address of Parmlist Byte Count
Fullword Boundary
↓ ↓

Byte 1 Byte 2 Byte 3 Byte 4

Number of bytes in following list
X'00' Beginning address of SORT or MERGE statement
Ending address of SORT or MERGE statement
Required in
Beginning address of RECORD statement
order shown
Ending address of RECORD statement
Address of E15 or E32 exit routine (Zeros if none)
Address of E35 exit routine (Zeros if none)
X'00' Main storage value
X'01' Reserved main storage value
X'02' Beginning address of MODS statement
Ending address of MODS statement
X'03' Beginning address of message DD name to replace SYS-
OUT
X'04' Number of input files (Use for merge with E32 exit only)

Optional X'05' Beginning address of DEBUG statement (Not processed)

Ending address of DEBUG statement
X'06' Beginning address of ALTSEQ statement
Ending address of ALTSEQ statement
X'07' Beginning address of SUM statement
Ending address of SUM statement
X'08' Beginning address of INCLUDE/OMIT statement
Ending address of INCLUDE/OMIT statement
Optional X'09' Beginning address of OUTREC statement
Ending address of OUTREC statement
X'0A' Beginning address of INREC statement
Ending address of INREC statement
X'0B' Beginning address of first OUTFIL statement
Ending address of first OUTFIL statement

Table 50. (Page 1 of 2) The 24-Bit Parameter List

6–6 Syncsort MFX Programmer’s Guide

 Byte 1 Byte 2 Byte 3 Byte 4
. .
. .
. .
X'0B' Beginning address of nth OUTFIL statement
Ending address of nth OUTFIL statement
X'0E' Beginning address of first JOINKEYS statement
Ending address of first JOINKEYS statement
X'0E' Beginning address of second JOINKEYS statement
Ending address of second JOINKEYS statement
X'0F' Beginning address of REFORMAT statement
Ending address of REFORMAT statement
X'10' Beginning address of JOIN statement
Ending address of JOIN statement
X'11' Beginning address of DUPKEYS statement
Ending address of DUPKEYS statement
X'F6' Beginning address of translation table
X'F7' User exit address constant
X'FD' IMS flag
X'FE' Pointer to STAE work area (May code zeros if none)
X'FF' Message option (Code in EBCDIC)
Optional DIAG option (Code in EBCDIC)
BALN, OSCL or POLY (Not processed)
CRCX, PEER or LIST (Not processed)
DD name (ddname) prefix to replace SORT in JCL (Code in EBCDIC)

Table 50. (Page 2 of 2) The 24-Bit Parameter List

Note: Empty boxes indicate the contents are immaterial and not examined.

Syncsort MFX Programmer’s Guide 6–7

 .
.
.
LA 1,POINTER Load address of pointer to parameter list
ATTACH EP=SORT Initiate MFX
BR 14 Return to invoking program
.
.
.
CNOP 2,4 Align to last 2 bytes of a word
BYTECNT DC Y(24) Byte count of parameter list
DC A(SORTBEG) Beginning address of sort statement
DC A(SORTEND) Ending address of sort statement
DC A(RECBEG) Beginning address of record statement
DC A(RECEND) Ending address of record statement
DC 2F'0' No E15 or E32 indicated. No E35.
POINTER DC X'80' Indicates parameter list's pointer
DC AL3(BYTECNT) Address of parameter list
SORTBEG DC C' SORT FIELDS=(1,16,CH,A),SKIPREC=100'
SORTEND DC C' '
RECBEG DC C' RECORD TYPE=F,LENGTH=80'
RECEND DC C' '

Figure 331. Sample Invoked Sort

In this example, only the required seven entries appear in the parameter list. The
invoked sort skips the first 100 records in SORTIN, a data set of 80-byte fixed-
length records, and sorts the remainder in ascending sequence according to the
character data in its first 16 bytes. Additional parameter-list entries may appear in
any order after these required seven; the contents of the first byte of the word
signals which optional parameter it is.
For invoked merge applications using the 24-bit parameter list, the number of
input files must be specified on either the X'04' entry or the FILES=n parameter on
the MERGE control statement. However, when using the 31-bit or 64-bit parameter
lists, the number of input files for a merge must be specified on the FILES=n
parameter.

Optional Parameters
Code in
Contents of Bytes 2 - 4
Byte 1
X'00' Indicates how much main storage MFX is to use. Code C'MAX' or a
hexadecimal number of bytes.
X'01' Indicates how much main storage should be reserved for data handling
by the invoking program during sort execution. MFX will use all avail-
able main storage less the hexadecimal number of bytes specified here.
This parameter takes precedence over the X'00' entry discussed above.

Table 51. (Page 1 of 3) Optional Parameters for the 24-Bit Parameter List

6–8 Syncsort MFX Programmer’s Guide

 Optional Parameters
Code in
Contents of Bytes 2 - 4
Byte 1
X'02' Gives the beginning address of a MODS control statement image. The
last 3 bytes of the next word must contain the ending address of this
image. All exits may be specified in the MODS statement image except
for E32 (use entry 6 of the parameter list for this exit). An E15/E35 exit
is specified in entries 6-7 or in the MODS image, but not in both.
X'03' Specifies the address of a replacement for the message data set’s DD
name. Eight characters are used for the name; the first character must
be alphabetic.
X'04' Specifies the hexadecimal number of SORTIN files-required whenever
an E32 exit supplies the input to the merge.
X'05' MFX accepts the DEBUG control statement parameter but does not
process it. Both this fullword (the beginning address of the DEBUG
statement) and the next (the ending address) are ignored.
X'06' Gives the beginning address of an ALTSEQ control statement image.
The last 3 bytes of the next word must contain the ending address of
this image.
X'07' Gives the beginning address of a SUM control statement image. The
last 3 bytes of the next word must contain the ending address of this
image.
X'08' Gives the beginning address of an INCLUDE/OMIT control statement
image. The last 3 bytes of the next word must contain the ending
address of this image.
X'09' Gives the beginning address of an OUTREC control statement image.
The last 3 bytes of the next word must contain the ending address of
this image.
X'0A' Gives the beginning address of an INREC control statement image.
The last 3 bytes of the next word must contain the ending address of
this image.
X'0B' Gives the beginning address of an OUTFIL control statement. The last
3 bytes of the next word must contain the ending address of this image.
This parameter may be specified more than once to accommodate mul-
tiple output file specifications.
X'0E' Gives the beginning address of a JOINKEYS control statement. The
last 3 bytes of the next word must contain the ending address of this
image. This parameter should be specified twice for the two
JOINKEYS specifications required in a join application.
X'0F' Gives the beginning address of a REFORMAT control statement
image. The last 3 bytes of the next word must contain the ending
address of this image.
X'10' Gives the beginning address of a JOIN control statement image. The
last 3 bytes of the next word must contain the ending address of this
image.

Table 51. (Page 2 of 3) Optional Parameters for the 24-Bit Parameter List

Syncsort MFX Programmer’s Guide 6–9

 Optional Parameters
Code in
Contents of Bytes 2 - 4
Byte 1
X'11' Gives the beginning address of a DUPKEYS control statement image.
The last 3 bytes of the next word must contain the ending address of
this image.
X'F6' Specifies the address of a 256-byte translation table, used to alter the
collating sequence. This parameter takes precedence over the ALTSEQ
control statement image.
X'F7' Optional user exit address constant which can be used to pass informa-
tion between the invoking program, an E15 exit, and/or an E35 exit.
MFX passes these 3 bytes to an E15 exit at offset 5 in an E15 parame-
ter list and to an E35 exit at offset 9 in an E35 parameter list. Offset 4
in the E15 parameter list and offset 8 in the E35 parameter list will
initially contain an X'00'.
X'FD' Only the first byte is processed, flagging this as an IMS-initiated sort,
processing variable-length records too short to contain all of the sort/
merge control field(s).
X'FE' If non-zero, these 3 bytes contain the address of a 104-byte STAE work
area.
X'FF' Indicates the MSG/FLAG PARM coding. For the MSG option, specify
AB, AC, AP, CB, NO, PC, SC or SP in bytes 3-4. For the FLAG option,
specify NOF, (I) or (U) in bytes 2-4. This parameter is coded in EBC-
DIC.
DIAG coded in EBCDIC in the entire word turns on the IOERR=ABE and RC16=ABE
PARM options.
BALN/OSCL/POLY coded in EBCDIC in the entire word is accepted but not processed by
MFX.
CRCX/PEER/LIST coded in EBCDIC in the entire word is accepted but not processed by
MFX.
xxxx with the first character alphabetic or national (do not use DIAG, PEER, CRCX, etc.)
represents the DD name prefix to be used in place of SORT in the JCL.

Table 51. (Page 3 of 3) Optional Parameters for the 24-Bit Parameter List

Return Codes
When the sort terminates, returning control to the calling program, it places a
return code in Register 15:
0 indicates normal termination;
16 indicates an unsuccessful sort.
The calling program typically tests the contents of Register 15, branching to the
normal-sort or sort-error end of job routine.

6–10 Syncsort MFX Programmer’s Guide

Sample Assembler Invocation Using 24-Bit Parameter List
.
.
.
LA 1,PTRWORD Load address of parameter list pointer
LINK EP=SORT Initiate MFX
LTR 15,15 Test return code
BNZ SORTERR Branch on error condition
B SORTOK Branch to normal processing
CNOP 0,4 Fullword alignment for pointer
PTRWORD DC X'80' Indicates pointer to parameter list
DC AL3(PARMS) Address of parameter list
DS H Unused first 2 bytes of first parameter
PARMS DC Y(PARMSEND-PARMSBEG) Byte count of remaining parameters
PARMSBEG DC A(SORTBEG) Beginning address of sort statement
DC A(SORTEND) Ending address of sort statement
DC A(RECBEG) Beginning address of record statement
DC A(RECEND) Ending address of record statement
DC F'0' No E15/E32 exit routine
DC F'0' No E35 exit routine
DC X'03' Indicates SYSOUT DD name change
DC AL3(MSGNAME) Address of SYSOUT DD name replacement
DC X'08' Indicates INCLUDE/OMIT parameter
DC AL3(OMITBEG) Beginning address of OMIT statement
DC A(OMITEND) Ending address of OMIT statement
DC X'07' Indicates SUM parameter
DC AL3(SUMBEG) Beginning address of SUM statement
DC A(SUMEND) Ending address of SUM statement
DC X'0B' Indicates OUTFIL parameter
DC AL3(OUTBEG1) Beginning address of first OUTFIL statement
DC A(OUTEND1) Ending address of first OUTFIL statement
DC X'0B' Indicates OUTFIL parameter
DC AL3(OUTBEG2) Beginning address of second OUTFIL statement
DC A(OUTEND2) Ending address of second OUTFIL statement
PARMSEND EQU * End of parameter list
SORTBEG DC C' SORT FIELDS=(1,20,A,35,8,A),' Begin SORT statement image
DC C'FORMAT=CH' Continue SORT statement image
SORTEND DC C' ' End SORT statement image
RECBEG DC C' RECORD TYPE=F' Begin RECORD statement image
RECEND DC C' ' End RECORD statement image
OMITBEG DC C' OMIT COND=(21,8,PD,EQ,0)' Begin OMIT statement image
OMITEND DC C' ' End OMIT statement image
SUMBEG DC C' SUM FIELDS=(21,8,PD)' Begin SUM statement image
SUMEND DC C' ' End SUM statement image
OUTBEG1 DC C' OUTFIL FILES=1,' Begin first OUTFIL statement image
DC C' HEADER1=(50X,''CHANGES'
DC C' TO W-2 FORMS'',//,'
DC C'50X,''JANUARY THROUGH JUNE'
DC C' 1993'')'
OUTEND1 DC C' ' End first OUTFIL statement image

Figure 332. (Page 1 of 2) Sample Assembler Invocation Using 24-Bit Parameter List

Syncsort MFX Programmer’s Guide 6–11

 OUTBEG2 DC C' OUTFIL FILES=2,' Begin second OUTFIL statement image
DC C'HEADER1=(19X,''EMPLOYEE'','
DC C'10X,''DEPARTMENT CODE'','
DC C'10X,''CHANGE''),'
DC C'OMIT=(29,4,PD,NE,0)' OMIT condition
OUTEND2 DC C' ' End second OUTFIL statement image
MSGNAME DC CL8'MESSAGES' SYSOUT DD name replacement
.
.
.
SORTERR DS 0H Error routine for unsuccessful sort
.
.
.
BR 14 Return
SORTOK DS 0H Normal processing for successful sort
.
.
.
BR 14 Return

Figure 332. (Page 2 of 2) Sample Assembler Invocation Using 24-Bit Parameter List

This example sorts fixed-length records by the character data in its first 20 bytes
and, where two records have identical data in this field, by the character data in
bytes 35-42; these fields are collated in ascending order. Note the continuation of
the SORT statement image using consecutive DC instructions. There is no special
significance to the break after the FIELDS parameter -- a control statement image
can be divided at any point in this way. The SORTIN file is edited by the OMIT
statement, which will eliminate any records with zero in bytes 21-28 before sorting
begins; these 8 bytes constitute the SUM field. MFX messages are written to the
data set specified by the MESSAGES DD name. Two OUTFIL parameters have
been specified, producing multiple output files. The first OUTFIL will receive data
from every sorted input record, producing a company-wide report. The second
OUTFIL will receive selected data only, as defined by the OMIT condition,
producing a departmental report.

The 31-Bit Extended Parameter List

The extended parameter list allows the sort to interface with invoking programs
that may require 31-bit addresses or which may use the 31-bit addressing mode
(AMODE).
Only the first word of the extended parameter list is required. The high order bit
must be zero to identify this as a 31-bit parameter list. The subsequent words of
this list are optional, and because there is no code in the high order byte, as in the
24-bit parameter list, their positional order must be maintained. Thus, when
coding the list be sure to code a fullword of zeros when omitting one of the optional

6–12 Syncsort MFX Programmer’s Guide

 parameters. The last parameter word specified in the list must be followed by the
4-byte field X'FFFFFFFF'.
The 31-bit parameter list has the following format:

REGISTER 1
d 31-bit address of start of parameter list
↓

Bit
Bits 1 through 31
0
Address of halfword containing the length of
Required +0 0
control statement images (zeros if none)
+4 m Address of user E15 or E32 (zeros if none)
+8 m Address of user E35 (zeros if none)
+12 User exits address constant (zeros if none)
Optional
+16 d Address of ALTSEQ translation table (zeros if
none)
+20 d Address of STAE area field (zeros if no STAE
routine)
+24 m Address of user exit E18 (zeros if none)
+28 m Address of user exit E39 (zeros if none)
+32 Call identifier (C 'nnnn')
Required +36 X'FFFFFFFF' (required)

Table 52. 31-Bit Extended Parameter List

Note: d indicates a bit is immaterial and not examined.
The following table provides an explanation of the contents of the extended
parameter list.

Syncsort MFX Programmer’s Guide 6–13

 Address Contents
+0 (Required) First word of the parameter list. The high order bit must be zero
to identify this as an extended parameter list. The other 31 bits contain the
address of a halfword which contains the length of the following control
statement images. A value of 0 (zero) represents a null list and control state-
ment images must be supplied through the $ORTPARM DD statement.
+4 (Optional) Address of the E15 or E32 exit routine. This address may point
anywhere in memory. The m bit (bit 0) means: 0=Enter the exit with 24-bit
addressing in effect (AMODE 24); 1=Enter the exit with 31-bit addressing in
effect (AMODE 31).
+8 (Optional) Address of the E35 exit routine. This address may point anywhere
in memory. The m bit (bit 0) means: 0=Enter the exit with 24-bit addressing
in effect (AMODE 24); 1=Enter the exit with 31-bit addressing in effect
(AMODE31).
+12 (Optional) User exit address constant which can be used to pass information
between the invoking program, an E15 exit routine, and/or an E35 exit rou-
tine. MFX passes these 4 bytes to an E15 exit routine at offset 4 in an E15
parameter list and/or to an E35 exit routine at offset 8 in an E35 parameter
list.
+16 (Optional) Address of ALTSEQ translation table. It can point anywhere in
memory and has a length of 256 bytes.
+20 (Optional) If non-zero, the address of a 112-byte STAE work area.
+24 (Optional) Address of the E18 exit routine. This address may point anywhere
in memory. The m bit (bit 0) means: 0=Enter the exit with 24-bit addressing
in effect (AMODE 24); 1= Enter the exit with 31-bit addressing effect
(AMODE 31).
+28 (Optional) Address of the E39 exit routine. This address may point anywhere
in memory. The m bit (bit 0) means: 0=Enter the exit with 24-bit addressing
in effect (AMODE 24); 1= Enter the exit with 31-bit addressing in effect
(AMODE 31).
+32 (Optional) Four displayable characters that will uniquely identify this partic-
ular call to MFX. (When this parameter is provided, MFX produces the
WER428I message and includes these four characters in the text. This mes-
sage facilitates correlation of message output when MFX is called multiple
times by the same program.)
+36 (Required) The last parameter word specified in the list must be
X'FFFFFFFF', which indicates end of list. If optional entries are omitted,
this end-of-list indicator is moved up to the word immediately after the last
specified parameter.

Table 53. Explanation of the Contents of the 31-Bit Extended Parameter List
Note: An optional parameter becomes required if a subsequent parameter is to
appear.

6–14 Syncsort MFX Programmer’s Guide

Return Codes
When the sort terminates, returning control to the calling program, it places a
return code in Register 15:
0 indicates normal termination;
16 indicates an unsuccessful sort.
The calling program typically tests the contents of Register 15, branching to the
normal-sort or sort-error end of job routine.
The following examples demonstrate how to code an extended parameter list. In
this example, all exits reside below the 16-megabyte line and should be called with
24-bit AMODE set, except the E35 exit, which should be called with 31-bit AMODE
set.

.
.
.
LA 1,XLIST Point at Parameter List
LINK EP=SORT Initiate MFX
.
.
.
XLIST DC A(CNTLCARD) Address of Control Card Images
DC A(E15EXIT) Address of E15 Exit
DC A(E35EXIT+X'80000000') Address of E35 Exit
DC F'0' User Address Constant
DC A(ALTSEQ) Address of ALTSEQ
Translation Table
DC A(STAE) Address of STAE Area Field
DC A(E18EXIT) Address of E18 Exit
DC A(E39EXIT) Address of E39 Exit
DC X'FFFFFFFF' End of Parameter List
CNTLCARD DC 0H
DC Y(CNTLLEN)
CNTLCRD2 DC C' SORT FIELDS=(1,16,CH,A)'
DC C' RECORD TYPE=F,LENGTH=80'
CNTLLEN EQU *-CNTLCRD2

Figure 333. Sample Invoked Sort with Both 24-bit AMODE & 31-bit AMODE Set

This next example demonstrates how to code an extended parameter list in which
all exits reside above the 16-megabyte line and should be called with AMODE 31
set.

Syncsort MFX Programmer’s Guide 6–15

 .
.
.
LA 1,XLIST Point at Parameter List
LINK EP=SORT Initiate MFX
.
.
.
XLIST DC A(CNTLCARD) Address of Control Card Images
DC A(E15EXIT+X'80000000') Address of E15 Exit
DC A(E35EXIT+X'80000000') Address of E35 Exit
DC F'0' User Address Constant
DC A(ALTSEQ) Address of ALTSEQ Translation Table
DC A(STAE) Address of STAE Area Field
DC A(E18EXIT+X'80000000') Address of E18 Exit
DC A(E39EXIT+X'80000000') Address of E39 Exit
DC X'FFFFFFFF' End of Parameter List
CNTLCARD DC 0H
DC Y(CNTLLEN)
CNTLCRD2 DC C' SORT FIELDS=(1,16,CH,A)'
DC C' RECORD TYPE=F,LENGTH=80'
CNTLLEN EQU *-CNTLCRD2

Figure 334. Sample Invoked Sort with 31-bit AMODE Set

6–16 Syncsort MFX Programmer’s Guide

Sample Assembler Invocation Using 31-Bit Parameter List
.
.
.
LOAD EP=E15EXIT Load E15 exit
ST R0,E15ADDR Store E15 AMODE+address
LA 1,XLIST Point at parameter list
LINK EP=SORT Initiate MFX
LTR R15,R15 Test return code
BNZ SORTERR Branch on error condition
B SORTOK Branch to normal processing
.
.
.
XLIST DC A(CARDLEN) Address of control statements
E15ADDR DC A(0) Address of E15 routine
DC A(0) No E35 routine
DC A(0) User exit address constant
DC X'FFFFFFFF' End of parameter list
CARDLEN DS 0H Control statement area
DC Y(CARDEND-CARDBEG) Length of character string
CARDBEG DC C'SORT FIELDS=(1,20,A,35,8,A),' Begin SORT image
DC C'FORMAT=CH' Continue SORT image
DC C'RECORD TYPE=F,LENGTH=80 ' RECORD image
DC C'OMIT COND=(21,8,PD,EQ,0) ' OMIT image
DC C'SUM FIELDS=(21,8,PD) ' SUM image
DC C'OUTFIL FILES=1,' First OUTFIL image
DC C'HEADER1=(50X,''CHANGES'
DC C' TO W-2 FORMS'',//,'
DC C'50X,''JANUARY THROUGH JUNE'
DC C'1992'')' End first OUTFIL image
DC C'OUTFIL FILES=2,' Second OUTFIL image
DC C'HEADER1=(19x,''EMPLOYEE'','
DC C'10X,''DEPARTMENT CODE'','
DC C'10X,''CHANGE'')'
DC C',OMIT=(29,4,PD,NE,0)' End second OUTFIL image
CARDEND EQU *
SORTERR DS 0H Error routine for unsuccessful sort
.
.
.
BR 14 Return
SORTOK DS 0H Normal processing for successful sort
.
.
.
BR 14 Return

Figure 335. Sample Assembler Invocation Using 31-Bit Parameter List

This example sorts fixed-length records by the character data in its first 20 bytes
and, where two records have identical data in this field, by the character data in
bytes 35-42; these fields are collated in ascending order. Note the continuation of
the SORT statement image using consecutive DC instructions. There is no special
significance to the break after the FIELDS parameter - a control statement image
can be divided at any point in this way. The SORTIN file is edited by the OMIT

Syncsort MFX Programmer’s Guide 6–17

 statement, which will eliminate any records with zero in bytes 21-28 before sorting
begins; these 8 bytes constitute the SUM field. Two OUTFIL parameters have been
specified, producing multiple output files. The first OUTFIL will receive data from
every sorted input record, producing a company-wide report. The second OUTFIL
will receive selected data only, as defined by the OMIT condition, producing a
departmental report.

The 64-Bit Parameter List

The 64-bit parameter list should be used when you have E15, E32, or E35 user
exits that run in AMODE64 and would like to pass records with 64-bit addresses
back to MFX. These exits can either reside in the invoking program or, for E15 and
E35, be in the load modules that are defined on the MODS statement using the
N64 link-editing code.
To use the 64-bit parameter list, you must use either SORT64 or ICEMAN64 as the
entry point name in the system macro that you use to invoke MFX. A 64-bit
Register 1 should point to the parameter list.
The parameter list is X’88’ long. It defines the same information as is defined in the
31-bit parameter list, with the addition of 64-bit addresses and exit AMODEs for
some fields. Except for the flag bit indicators, the entries are each 8 bytes, which
allows for 64-bit addresses for some entries. The MFX control statement images
and an optional ALTSEQ translation table can reside above the bar (that is, have a
64-bit address). The other entries have binary zeros in the first 4 bytes.
Table 54 provides an explanation of the contents of the extended parameter list. All
bits and bytes indicated as reserved should be set to zero when creating the
parameter list. All full 8-byte addresses in the list are either a 64-bit address, or a
24 or 31-bit address left-padded with binary zeros. For the exit AMODE flag bits,
exactly one of the bits should be set for each exit whose address is specified in the
parameter list. The exit AMODE flag bits do not apply to exits which are defined in
a MODS statement.

6–18 Syncsort MFX Programmer’s Guide

 Displacement
Length Bit Field
Hex Decimal
0 0 8 C’PL64SORT’ (identifier)
8 8 1 X’80’ AMODE 24 for E15 Exit or E32

X’40’ AMODE 31 for E15 Exit or E32

X’20’ AMODE 64 for E15 Exit or E32

X’10’ AMODE 24 for E35 Exit

X’08’ AMODE 31 for E35 Exit

X’04’ AMODE 64 for E35 Exit

X’02’ AMODE 24 for E18 Exit

X’01’ AMODE 31 for E18 Exit

9 9 1 X’80’ Reserved

X’40’ AMODE 24 for E39 Exit

X’20’ AMODE 31 for E39 Exit

X’10’ Reserved

X’08’ 1: E15/E32 64-bit Parameter List

0: E15/E32 31-bit Parameter List

X’04’ 1: E35 64-bit Parameter List

0: E35 31-bit Parameter List

X’02’ Reserved

X’01’ Reserved
A A 13 Reserved
X’80’ 1: Using Blocked E15 Exit

X’40’ 0: Using Blocked E35 Exit

17 23 1
X’20’ Reserved

X’01’ Reserved
Address of halfword containing length of
18 24 8 control statement images, followed by
images (zeros if none)
F’0’, Address of user exit E15 or E32
20 32 8
(zeros if none)
28 40 8 F’0’, Address of user exit E35 (zeros if none)

Table 54. (Page 1 of 2) The Extended Parameter List

Syncsort MFX Programmer’s Guide 6–19

 Displacement
Length Bit Field
Hex Decimal
F’0’, Address of user exit address constant
30 48 8
(zeros if none)
Address of ALTSEQ translation table
38 56 8
(zeros if none)
40 64 8 F’0’, Address of ESTAE area (zeros if none)
48 72 8 F’0’, Address of user exit E18 (zeros if none)
50 80 8 F’0’, Address of user exit E39 (zeros if none)
58 88 8 F’0’, Call identifier (C’nnnn’) (zeros if none)
64-bit Address of user block list parameter
60 96 8
area used by blocked E15 and/or E35
68 104 32 Reserved

Table 54. (Page 2 of 2) The Extended Parameter List

6–20 Syncsort MFX Programmer’s Guide

 Table 55 provides an explanation of the E15 and E35 user exit parameters
available in the 64-bit parameter list, which support the transfer of blocks of
records between MFX and the E15/E35 user exits.
Displacement
Length Bit Field
Hex Decimal
Block List Parameters for E15 Exit
0 0 8 D Address of User Block List Area
Maximal length of blocks with records at
8 8 8 D each return to SORT (for Block List Type 2
only)
Maximal number of blocks with records at
10 16 8 D each return to DFSORT (for Block List
Type 1 or Type 2)
18 24 4 F Maximal length of record
1C 28 1 C Format of record (F-FLR/V-VLR)
1D 29 1 X Block List Type for 15 exit (1/2/...)
1E 30 2 XL2 Reserved
Block List Parameters for E35 Exit
0 0 8 D Address of User Block List Area
Maximal length of blocks with records at
8 8 8 D each return to SORT (for Block List Type 2
only)
Maximal number of blocks with records at
10 16 8 D each return to DFSORT (for Block List
Type 1 or Type 2)
18 24 4 F Maximal length of record
1C 28 1 C Format of record (F-FLR/V-VLR)
1D 29 1 X Block List Type 1 for 35 exit (1/2/...)
1E 30 2 XL2 Reserved

Table 55. Block List Parameters for E15/E35 User Exits

Syncsort MFX Programmer’s Guide 6–21

Syncsort Java Class
The Syncsort() Java class provides a wrapper around the Syncsort MFX sort/
merge/copy utility. When invoked from Java, the input data can be taken from
either existing data sets or produced directly from a Java application. Likewise, the
output data from Syncsort MFX can be written to a data set or read directly back
into the Java application. When writing from or reading into a Java application,
the java.ioInputStream and java.io.OutputStream can be used to read and
write data from the application to Syncsort MFX.

Constructor Detail
public Syncsort()

Method Summary

Modifier & Type Method Description

Void addAllocation Add an allocation request to be
(java.lang.String
sent to the Syncsort MFX
allocation)
invocation.
Void addControlStatement Add a control statement to be
(java.lang.String
sent to the Syncsort MFX
controlStatement)
invocation.
Void disableInputStream() Indicate that the records to
Syncsort MFX are to be
provided via the SORTIN DD
and not from Java.
Void disableOutputStream() Indicate that the output records
from Syncsort MFX are to be
sent to the SORTOUT DD and
not to Java.
java.io.OutputStream getChildStdinStream() Returns a java.io.OutputStream
that can be used to send
records to the child Syncsort
MFX process.
java.io.InputStream getChildStdoutStream() Returns a java.io.InputStream
that can be used to read output
records from the child Syncsort
MFX process.

6–22 Syncsort MFX Programmer’s Guide

 Modifier & Type Method Description
Int getReturnCode() Wait for Syncsort MFX to
complete.
java.util.List getStderrLines() Return a list of stderr output
lines (Strings) from the
spawned Syncsort MFX
process.
Void setInputStreamHasRdws() Indicate that the records to be
sent as input to Syncsort MFX
are variable length, and that
each record will prefixed by a
4-byte (big endian) RDW.
Void setInputStreamRecLen Set a fixed record length for
(int reclen)
the records that will be sent as
input to Syncsort MFX.
Void setOutputStreamHasRdws() Indicate that the records sent
from Syncsort MFX to java are
variable length, and that each
record will prefixed by a 4 byte
(big endian) RDW.
Void setOutputStreamRecLen Set a fixed record length for
(int reclen)
the records that will be
received from Syncsort MFX.
Void setSameAddressSpace Sets whether the child Syncsort
(boolean same)
MFX process will be spawned
in the same address space as
the Java JVM, or in a separate/
new BPX address space.

Syncsort MFX Programmer’s Guide 6–23

Methods Inherited from Class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, toString, wait

Method Detail
addAllocation
public void addAllocation(java.lang.String allocation)
Add an allocation request to be sent to the Syncsort MFX invocation.
The allocation strings are expected to be in the syntax required by BPXWDYN.
Typically, Syncsort MFX may have allocations for the following DDs:
• SORTIN – where Syncsort MFX gets sort/merge input.
• SORTOUT – where Syncsort MFX writes sort/merge output
Example:
Syncsort syncsort = new Syncsort(); syncsort.addAllocation
("alloc fi(sortin) da('HLQ.MY.DATASET') reuse shr msg(2)");
Note: Allocation requests must be added before calling execute() in order to have
any effect.

addControlStatement
public void addControlStatement(java.lang.String controlStatement)
Add a control statement to be sent to the Syncsort MFX invocation.
Example:
Syncsort syncsort = new Syncsort(); syncsort.addControlStatement("SORT
FIELDS=(1,80,CH,A)");
Note: Control statements must be added before calling execute() in order to have
any effect.

setInputStreamRecLen
public void setInputStreamRecLen(int reclen)
Set a fixed record length for the records that will be sent as input to Syncsort MFX.
This length must match the record length expected by Syncsort MFX as specified
by the RECORD control statement.
This method is mutually exclusive with setInputStreamHasRdws() and
disableInputStream().
Note: This property must be changed before calling execute() in order to have any
effect.
Parameters:
reclen – int the input record length for fixed records

6–24 Syncsort MFX Programmer’s Guide

setInputStreamHasRdws
public void setInputStreamHasRdws()
Indicate that the records to be sent as input to Syncsort MFX are variable length,
and that each record will prefixed by a 4-byte (big endian) RDW. This RDW value
must be less than or equal to the maximum record length expected by Syncsort
MFX as specified by the RECORD control statement. The RECORD control
statement must indicate that Syncsort MFX is to expect variable length records.
Note: This property must be changed before calling execute() in order to have any
effect.
This method is mutually exclusive with setInputStreamRecLen(int) and
disableInputStream().

disableInputStream
public void disableInputStream()
Indicate that the records to Syncsort MFX are to be provided via the SORTIN DD
and not from Java.
Note: This property must be changed before calling execute() in order to have any
effect.
This method is mutually exclusive with setInputStreamRecLen(int) and
setInputStreamHasRdws().

setOutputStreamRecLen
public void setOutputStreamRecLen(int reclen)
Set a fixed record length for the records that will be received from Syncsort MFX.
This length must match the record length produced by Syncsort MFX as specified
by the RECORD control statement.
Note: This property must be changed before calling execute() in order to have any
effect.
This method is mutually exclusive with setOutputStreamHasRdws() and
disableOutputStream().
Parameters:
reclen – int the output record length for fixed records.

setOutputStreamHasRdws
public void setOutputStreamHasRdws()
Indicate that the records sent from Syncsort MFX to Java are variable length, and
that each record will prefixed by a 4-byte (big endian) RDW. This RDW value will
be less than or equal to the maximum record length as specified by the RECORD

Syncsort MFX Programmer’s Guide 6–25

 control statement. The RECORD control statement must indicate that Syncsort
MFX is to produce variable length records.
Note: This property must be changed before calling execute() in order to have any
effect.
This method is mutually exclusive with setOutputStreamRecLen(int) and
disableOutputStream().

disableOutputStream
public void disableOutputStream()
Indicate that the output records from Syncsort MFX are to be sent to the
SORTOUT DD and not to Java.
Note: This property must be changed before calling execute() in order to have any
effect.
This method is mutually exclusive with setOutputStreamRecLen(int) and
setOutputStreamHasRdws().

getChildStdinStream
public java.io.OutputStream getChildStdinStream()
Returns a java.io.OutputStream that can be used to send records to the child
Syncsort MFX process. This data stream is used to provide input to Syncsort MFX
via a sort input exit as an alternative to DD:SORTIN.The data sent to this stream
should be fixed length records if setInputStreamRecLen(int)() has been used, or
RDW-prefixed records if setInputStreamHasRdws() has been used.
If however, disableInputStream() was set (the default), then this stream has no use,
as input to Syncsort MFX will come from DD:SORTIN. For better performance, this
OutputStream should be wrapped in a BufferedOutputStream.
Returns:
java.io.OutputStream
Throws:
java.lang.IllegalStateException – if called before execute()

getChildStdoutStream
public java.io.InputStream getChildStdoutStream()
Returns a java.io.InputStream that can be used to read output records from the
child Syncsort MFX process. This data stream will contain sorted output data from
Syncsort MFX via a sort output exit as an alternative to DD:SORTOUT. The data
sent to this stream should be fixed length records if setOutputStreamRecLen(int)()
has been used, or RDW-prefixed records if setOutputStreamHasRdws() has been
used.

6–26 Syncsort MFX Programmer’s Guide

 If however, disableOutputStream() was set (the default), then this stream has no
use, as output from Syncsort MFX will go to DD:SORTOUT.
For better performance, this InputStream should be wrapped in a
BufferedInputStream.
Returns:
java.io.InputStream
Throws:
java.lang.IllegalStateException – if called before execute().

getReturnCode
public int getReturnCode()
Wait for Syncsort MFX to complete. This method should be called after execute()
and after I/O from the child’s stdin and/or stdout has been processed.
Returns:
int – the Syncsort MFX return code. Value unpredictable if called before execute().
Throws:
RcException –if the child process terminates abnormally without a return code.

getStderrLines
public java.util.List getStderrLines()
Return a list of stderr output lines (Strings) from the spawned Syncsort MFX
process. This method only applies if isSameAddressSpace() = false. Otherwise,
Syncsort MFX messages will be written to Java’s stderr.
Returns:
List – a list of strings.

setSameAddressSpace
public void setSameAddressSpace(boolean same)
Sets whether the child Syncsort MFX process will be spawned in the same address
space as the Java JVM, or in a separate/new BPX address space.
Parameters:
same – flag to indicate whether the Syncsort MFX child invocation process should
run in the same address space as the JVM. If true, run the Syncsort MFX child
invocation process in the same address space as the JVM. If false, spawn a new
address space.

Syncsort MFX Programmer’s Guide 6–27

Java Wrapper Example

import com.ibm.jzos.*;
import java.io.BufferedInputStream;
import java.util.*;
import com.syncsort.zos.SyncSort;
public class dfexec1
{
public static void main(String args) throws Exception {
sort("<INPUT DSNAME>");
}
private static void sort(String inDs) throws Exception {
int key = 20;

SyncSort syncsort = new SyncSort();

syncsort.addAllocation("alloc fi(SORTIN) da(" + inDs + ") reuse shr msg(2)");
syncsort.setOutputStreamRecLen(80);
syncsort.addControlStatement("SORT FIELDS=(1," + key + ",CH,A)");

long startTime = System.currentTimeMillis();

syncsort.execute();
BufferedInputStream bis = new BufferedInputStream(syncsort.getChildStdoutStream());
byte buffer = new byte 80 ;
int bytesRead = 0;
while ((bytesRead = bis.read(buffer)) != -1) {
System.out.println(new String(buffer, 0, bytesRead));
}
int rc =0;
try {
rc = syncsort.getReturnCode();
} catch (RcException rce) {
System.out.println("Caught RcException: " + rce.getMessage());
rc = -1;
}
if (rc != 0) {
List stderrLines = syncsort.getStderrLines();
for (Iterator i=stderrLines.iterator(); i.hasNext(); ) {
System.err.println(i.next());
}
}
}
}

6–28 Syncsort MFX Programmer’s Guide

Chapter 7 The Coding and Use of Exit
Programs

What Is an Exit?
The term program exits refers to the various points in the sort program’s
executable code at which control can be passed to a user-written routine. Most exit
routines take control once for every record being processed, increasing overall
execution time and consuming main storage that would otherwise be used by the
sort. Exits should only be coded for tasks which cannot be accomplished with MFX
control statements.
Program exits are not allowed to take their own OS or VS checkpoints.
Program exits are labeled with a 2-digit decimal number, e.g., E35. Except for E61,
the first digit (1, 2 or 3) refers to the sort/merge phase at which the routine will get
control; an E61 routine can take control in Phase 1 or Phase 3. The second digit
refers to the number of that exit within the phase. Whenever possible, control
passes directly from Phase 1 to Phase 3, skipping the intermediate merge phase
and its associated exits: E21, E25 and E27.
As indicated in the following chart, the nature of the task determines the program
exit to be used.

Syncsort MFX Programmer’s Guide 7–1

 PHASE 1 PHASE 2 PHASE 3

E11 E14 E15 E16 E17 E18 E61 E21 E25 E27 E31 E32 E15 E35 E37 E38 E39 E61
TASK *

Prepare for
other exit X X X
routines

Create
input
records for
X X
sort (Phase
1) and copy
(Phase 3)

Create
input
X
records for
merge

Add records X X X

Delete
X X X X X
records

Change
X X X X X
records

Sum
X X X
records

Choose
action if
intermedi- X
ate storage
insufficient

Close other
exit data X X X
sets

Process
X X
read errors

Process
X
write errors

Check
X X X
labels

Table 56. (Page 1 of 2) Program Exits and Processing Phases

7–2 Syncsort MFX Programmer’s Guide

 PHASE 1 PHASE 2 PHASE 3

Modify a
collating X X
sequence

* E15 in Phase 3 for copy only

Table 56. (Page 2 of 2) Program Exits and Processing Phases

Loading the Exit Routines into Main Storage

The MODS statement identifies the exits to be taken and indicates the name of the
separately compiled, user-written routine to take control at that point. The same
routine (e.g., deleting selected records) could take effect in different phases, but
cannot be loaded more than once in a single phase.
Note that merge and copy are executed entirely in Phase 3 and are therefore
restricted in the exits which they can use. A merge application cannot use exits E11
through E27. A copy application can use exit E15 but not exits E32 or E61.
Assemble each routine as a separate program and place it in a partitioned data set
or in the SYSIN input stream; MFX copies the SYSIN routines to the SORTMODS
library for linkage editing. (If a SYSIN module is to be used at more than one exit
point, each exit must have its own compiled copy of the module in SYSIN.) If MFX
linkage edits an exit routine, the module must have an entry point whose name is
that of the MFX exit; for example, in order to function as an E35 routine, MYEXIT
must include an entry point or CSECT labeled E35.
If a routine has already been link-edited, this can be indicated in the MODS
statement. When all the exits in a particular phase need to communicate with one
another, the MODS statement can be used to instruct the sort to link-edit them
together.

Exit Conventions
The following conventions must be observed when using exits.
• Exits provided via the MODS control statement will be entered in the
addressing mode indicated by the linkage editor module attributes. Any exit
linkage-edited by MFX will be entered in 24-bit addressing mode, except a
separately linkage-edited exit E11, E21, or E31, which will be entered in the
mode set by the compiler or assembler when the module was compiled or
assembled.

Syncsort MFX Programmer’s Guide 7–3

 • Exit addresses provided via the 24-bit invoking parameter list format will be
entered in the 24-bit address mode.
• Exit addresses provided via the 31-bit invoking parameter list will be entered
in the address mode indicated in the exit address field. That is, if bit 0 of the
exit address is 0, the exit is entered in 24-bit mode; if bit 0 of the exit address is
1, the exit is entered in 31-bit mode.
• Exit addresses provided via the 64-bit parameter list will be entered in the
addressing mode indicated by the flag bytes in the parameter list.
• User exits may return to the sort in either 24- or 31-bit address mode.
• If an exit was entered in 24-bit address mode using a 32-bit exit parameter list,
the addresses passed to it will be 24-bit values that have a clean high-order byte
containing binary zeros (X'00'). Addresses returned to the sort must also be 24-
bit values with a high-order byte containing X'00' even though the exit could
return to the sort in the 31-bit mode.
• An exit entered in the 31-bit addressing mode using a 32-bit exit parameter list
may return an address containing a full 31-bit value. Users intending to pass
only a 24-bit address must therefore make sure that the address returned has
X'00' in the high-order byte. Failure to do so can have unpredictable results.
Note that certain addresses within parameter lists are still explicitly restricted
to 24-bit values. For example, E18 exit return parameter lists must consist of
fullword entries that are 1-byte codes and 3-byte addresses.
• An exit that was called by MFX with a 64-bit parameter list may return a 64-bit
record address. If such exits intend to pass only a 31-bit or 24-bit address back
to MFX, they should ensure that the high order bits of the 64-bit Register 1 are
zeros. This applies regardless of the addressing mode in which the exit was
called.

Register Conventions
The standard operating system conventions apply to register usage. Exit routines
must save and restore Registers 0 and 2-14. The sort/merge places these contents
in Register 1 and 13-15 for use by the exit routine when it takes control.
Register 1 The address of an MFX parameter list.
Register 13 The address of a register save area. For exits called with the
24-bit or 32-bit parameter list, the save area is 19 words,
where the first 18 words are a Format 0 save area that can be
used to save registers, and the 19th word can be used to pass
information between Assembler exits. For exits called with
the 64-bit exit parameter list, the save area is a 36 word For-

7–4 Syncsort MFX Programmer’s Guide

 mat 4 save area. The Format 0 and Format 4 save areas are
defined by the IHASAVER Assembler macro.
Register 14 MFX’s return address, in the low-order address bits of the
register. The high-order bit(s) may have undefined contents.
Register 15 The address of the entry point of the exit routine, in the low-
order address bits of the register. The high-order bit(s) may
have undefined contents.

The Exit Communication Area

When an exit routine is given control using a 32-bit exit parameter list, Register 13
points to a 19-word area. The 19th word of this area can be used to pass
information between Assembler exits. For example, when the COMMAREA PARM
is used, the 19th word can be set to point to the exit communication area
COMMAREA provides. The first 2 bytes of this communication area give the length
of the area. You are free to change the entire communication area, including the
initial halfword.

REGISTER 13 * Address of communication area

↓ ↓
18-word save area *

↓
00 04 LIST

Figure 336. User Communication Area for Assembler Exit Using COMMAREA PARM

For COBOL or C exits, the address and length of this area are passed in the
COBOL or C program’s parameter list. In this case, there is no halfword preface –
the address points directly to the communication area.

Syncsort MFX Programmer’s Guide 7–5

Exits E11, E21, and E31 - Preparing for Other Exit Routines
These exits are unusual in that they are entered only once, at the beginning of their
associated phase. Because of this, they may be separately link-edited and are
efficiently used to prepare for other exit routines (e.g., to open files or initialize
variables). There are no parameter lists or return codes for these exits.

Exit E32 - Invoked Merge Only: Creating Input Records

This exit can only be used for an invoked merge and must be coded in line with the
invoking program. It therefore never appears on the MODS statement. When an
E32 routine is used, all SORTINnn DD statements will be ignored by the merge;
the exit must supply all the input records, and the number of input files to be
created must be supplied by either the invoking program’s parameter list or the
FILES=n parameter on the MERGE control statement.
Whenever the merge requires a new input record, MFX calls the E32 routine,
passing it the address of a parameter list with three entries in Register 1.
The 32-bit parameter list has the following format:

PARAMETER LIST
1) Word 1: Number of next input file
2) Word 2: Address of the next input record (set by exit)
3) Word 3: User exit address constant

Figure 337. 32-bit Parameter List for E32

The 64-bit E32 parameter list can be enabled from the 64-bit sort invocation parameter list. It
has the following format:

PARAMETER LIST
1) Word 1: X’00000000’ Word 2: Number of next input file
2) Words 3 and 4: 64-bit address of next input record (set by exit)
3) Word 5: X’00000000’ Word 6: User exit address constant

Figure 338. 64-bit Parameter List for E32

The first entry of the parameter list contains a hexadecimal representation of the
input file MFX is currently processing. This is initialized as 0 for the first file and

7–6 Syncsort MFX Programmer’s Guide

 incremented by 4 every time a new SORTINnn file is to be accessed. When the E32
encounters end-of-file on a SORTINnn file it should return RC=8 to MFX, which
will no longer request input from that file (i.e., that input file number).
The E32 routine must respond to three different cases: (a) MFX already has all the
input records; (b) the previous record finished an input file; and (c) there is at least
one more record to be added to the file with this file number. Only in the last case
will the E32 supply a record address to the merge, placing it in the second entry of
the parameter list. The E32 also places the appropriate return code in Register 15.

Return Codes
8 End of file. This tells MFX that a particular file has been completed
and to make no further request for records from that file.

12 Insert this record. This tells MFX to accept a new record from the
input file requested.

16 End of merge. This terminates MFX with a critical error.

Exit E14 - Deleting, Summing, Changing Records

Exit E14 may be used to change the contents of data fields, or to delete or sum
records during Phase 1. Unlike an E15 exit routine, it cannot be used to add
records. An E14 exit program requires:
• at least one SORTWKxx data set, assigned to disk;
• fixed-length input records.
This exit is given control whenever MFX is about to add a record to an output
sequence. Since it does not take control before the first record of that sequence, the
routine always has access to a pair of sequenced records (e.g., for summation
purposes). MFX passes the exit program a two-word parameter list by loading its
address into Register 1. The exit must not destroy the contents of this parameter
list. The first word, which is on a fullword boundary, contains the address of the
record about to be placed in an output sequence; the second word contains the
address of the record that has just been put into the output buffer.

PARAMETER LIST
Word 1: Address of record leaving Phase 1
Word 2: Address of the latest record in output buffer

Figure 339. Parameter List for E14

Syncsort MFX Programmer’s Guide 7–7

 There are two constraints on the type of processing the exit may accord this record
pair:
• Sort control fields should not be changed since this may cause an out-of-
sequence condition.
• If a record is to be changed, it should first be moved to a work area.
After record processing is completed, the exit routine must place the appropriate
return code in Register 15. The exit must save and restore all registers except those
used in linking to the sort/merge.

Return Codes
0 Accept this record. This instructs MFX to accept the record whose
address is in the first word of the parameter list and place it in the
output buffer. The exit must also load the (work area) address of this
record into Register 1 before returning control to the sort.
4 Delete this record. This instructs MFX to delete the record whose
address is in the first word of the parameter list. Do not place the
address of this record in Register 1. This return code might be
employed, for example, after using this record to update the previous
(output) record. Assuming this does not complete an output sequence
for Phase 1, the next execution of the E14 will find the same address
in the second word of the parameter list.

Exit E15 - Creating, Revising, or Analyzing the Input File

Where an input data set already exists, this exit is used to add, delete and/or
change input records. This exit is also used to analyze SORTIN via HISTOGRM (a
HISTE15 application) or to create the entire input file. It can be used when sorting
or copying records.
When used in conjunction with an input file, this exit is given control every time a
record is brought into Phase 1 of a sort or Phase 3 of a copy. In passing control to
the E15 exit routine, MFX places the address of a parameter list in Register 1. This
parameter list contains two entries.
For 32-bit parameter lists, each entry is one word. In the first word, the first byte
contains X'00'; the last 3 bytes contain the address of the next record to be
processed. The first word contains a zero address when there is no such record (i.e.,
when SORTIN end-of-file is reached or when the input data set is empty).
Word 2 contains the user address constant. On the initial call to the E15 exit, it will
contain the value specified in the invoking parameter list. If this value was
specified in a 24-bit invoking parameter list, it will have the high-order byte set to
X'00'. If the value was omitted or MFX was JCL invoked, Word 2 will contain

7–8 Syncsort MFX Programmer’s Guide

 binary zeros. This word may be changed by the E15 exit whenever it is entered. If
used in a sort application, the value will be returned on the subsequent call to the
E15. If used in a copy application and an E35 is present, the value on the
subsequent call to the E15 will reflect any modification made to the User Address
Constant by the E35. In a sort application, the initial entry to the E35 will contain
the value last returned from the E15.

PARAMETER LIST
1) Word 1: Address of the new record
2) Word 2: User exit address constant

Figure 340. 32-bit Parameter List for E15

The 64-bit E15 parameter list can be enabled from the 64-bit sort invocation
parameter list. See p. 7.4 for information on returning records to MFX in register 1
and see ”Register Conventions” on p 7.4 for a description of the register 13 save
area for exits called with the 64-bit exit parameter list. The same information is
passed to the exit as for the 32-bit parameter list, but double-word entries are used:

PARAMETER LIST
1) Word 1: X’00000000’ Word 2: Address of the new record
2) Word 3: X’00000000’ Word 4: User exit address constant

Figure 341. 64-bit Parameter List for E15

E15 record processing has these two constraints:

• If a record is to be changed, it should first be moved to a work area.
• When the input data set consists of variable-length records, the first 4 bytes
must contain the Record Descriptor Word, giving the length of the record.
When the program has finished processing the record, it must place the
appropriate return code in Register 15.

Coding the E15 Exit Routine for an Invoked Sort or Copy

When MFX is initiated from an ATTACH, LINK or XCTL macro, there are two
ways to include an E15 exit routine: (1) code the E15 exit routine in line with the
invoking program and specify the address of its entry point in the appropriate
entry of the parameter list; or (2) define the separately compiled routine in the
MODS control statement. When the exit routine is coded in line with the calling
program, it must supply the entire input data set; MFX will ignore a SORTIN DD
statement, if present. Data set creation is done by supplying the sort with one
record at a time, placing its address in Register 1 and a return code of 12 in
Register 15. After the last record has been submitted, the exit passes a return code
of 8.

Syncsort MFX Programmer’s Guide 7–9

Return Codes
0 Accept this record. This instructs MFX to accept the record the exit
has just examined. Place the (work area) address of this record in
Register 1. Exits using the 32-bit parameter list return a 31-bit
address; exits using the 64-bit parameter list return a 64-bit address.
This return code is used when selectively editing records from an
input file; it passes the (possibly altered) record back to the sort. The
RECORD statement is required if the exit routine changes the maxi-
mum record length; code the old maximum length as l1, the new max-
imum as l2.
4 Delete this record. MFX will delete the record just examined. There is
no need to load the address of this record into Register 1.
8 Do not return to this exit. This instructs MFX to close the exit for the
remainder of the sort application. This return code might be used at
SORTIN end-of-file (signalled by a zero address in the parameter
list) to indicate that extra records will not be added at this point.
There is no need to load a record address into Register 1 when
passing a return code of 8. If SORTIN is present, the current input
record and all subsequent records will be processed by MFX.
12 Insert a record. This tells MFX that the exit routine has located a
record which should be added to the input data set before the record
whose address appears in the parameter list. Load the address of the
new record into Register 1. Exits using the 32-bit parameter list
return a 31-bit address; exits using the 64-bit parameter list return a
64-bit address. When MFX returns control to the E15, the parameter
list will be unchanged. The exit routine can then add another record
or process the current one.
This return code can be used to add records to the end of the input
data set or to create the entire input data set. MFX returns to the
exit routine, adding records without changing the parameter list (in
these cases, a zero address) until a different return code (i.e., RC=8)
is passed. When the input data set is created in this way, the
RECORD statement is required and must specify both TYPE and
LENGTH.
16 Terminate MFX. This tells MFX to terminate and return to the call-
ing program or the supervisor. MFX uses a completion code of 16 to
indicate that the sort was unsuccessful.

7–10 Syncsort MFX Programmer’s Guide

Coding a COBOL E15 Exit Routine
A COBOL E15 exit program can be indicated through the EXEC statement’s PARM
option (PARM='E15=COB'), the MODS control statement or the $ORTPARM DD
statement.
Like any other E15 exit routine, the COBOL E15 exit routine is called each time a
record is brought into Phase 1 of a sort or Phase 3 of a copy. Communication
between MFX and the COBOL exit takes place in the LINKAGE SECTION of the
COBOL program. For example, records are passed to the COBOL routine in the
second definition (RECORD-UP) area of the LINKAGE SECTION.
If the COBOL exit routine uses any verb (EXHIBIT, DISPLAY, TRACE) which
results in output to the SYSOUT DD statement, there is a potential conflict with
MFX’s use of this DD statement. It is therefore recommended that you separate the
output by using either MFX’s MSGDD PARM option or the COBOL compiler’s
SYSx parm.

The LINKAGE SECTION

The LINKAGE SECTION examples that follow show the parameters required for
passing fixed-length and variable-length records to the sort. The data-names and
conditional names used in the examples are arbitrary but each definition is
required. The complete programs from which the examples are taken follow the
discussion of the exit.

Syncsort MFX Programmer’s Guide 7–11

Example 1: Fixed-Length Records
LINKAGE SECTION.
01 EXIT-STATUS PIC 9 (8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.
01 RECORD-UP.
07 FILLER PIC 9(6).
07 R-SEQ2 PIC 9(2).
07 FILLER PIC X(92).
01 WORK PIC X(100).
01 DUMMY1 PIC X.
01 DUMMY2 PIC X.
01 DUMMY3 PIC X.
01 DUMMY4 PIC X.
01 DUMMY5 PIC X.

01 COMM-LEN PIC 9(4) COMPUTATIONAL.

01 COMMUNICATION-AREA.

05 COMM-AREA OCCURS 1 TO 256 TIMES

DEPENDING ON COMM-LEN PIC X.

Figure 342. Sample Fixed-Length Record

• For the first definition (EXIT-STATUS) specify PIC 9(8) COMPUTATIONAL.

(This area defines exit status codes.) When using 88 levels to define the exit
status codes, specify values 00, 04, and 08.
• For the second definition (RECORD-UP) define the SORTIN record.
• For the third definition (WORK) define the record that will be passed to MFX.
(This is the "work area.")
• For the fourth through the eighth definitions define dummy areas.
• For the ninth definition (COMM-LEN) specify PIC 9(4) COMPUTATIONAL.
This area defines the communication area length.
• For the tenth definition (COMMUNICATION-AREA) code an OCCURS clause
DEPENDING ON data-name PIC X.

7–12 Syncsort MFX Programmer’s Guide

Example 2: Variable-Length Records
LINKAGE SECTION.
01 EXIT-STATUS PIC 9 (8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.
01 RECORD-UP.
05 RU OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-RU PIC X.
01 WORK.
05 WK OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-WK PIC X.
01 IN-BUF PIC X(100).
01 DUMMY PIC X(4).
01 LEN-RU PIC 9(8) COMPUTATIONAL.
01 LEN-WK PIC 9(8) COMPUTATIONAL.
01 LEN-IB PIC 9(8) COMPUTATIONAL.
01 COMM-LEN PIC 9(4) COMPUTATIONAL.
01 COMMUNICATION-AREA.
05 COMM-AREA OCCURS 1 TO 256 TIMES
DEPENDING ON COMM-LEN PIC X.

Figure 343. Sample Variable-Length Record

• For the first definition (EXIT-STATUS) specify PIC 9(8) COMPUTATIONAL.

(This area defines exit status codes.)
• For the second definition (RECORD-UP) code an OCCURS clause with the
DEPENDING ON data-name option specifying (1) The minimum and
maximum number of bytes the variable SORTIN records contain (do not include
4 bytes for the RDW) and (2) DEPENDING ON data-name PIC X. Data-name is
defined in the sixth definition in the LINKAGE SECTION.
• For the third definition (WORK) code an OCCURS clause with the
DEPENDING ON data-name option specifying (1) The minimum and
maximum number of bytes for variable-length records to be passed to MFX (do
not include 4 bytes for the RDW) and (2) DEPENDING ON data-name PIC X.
Data-name is defined as the seventh definition in the LINKAGE SECTION.
• For the fourth definition specify a dummy level 01 data-name of any number of
bytes. (IN-BUF is the data-name used in this example.) Note that the level 01
data-name, used here as a dummy address, has no effect on the E15 routine for
variable-length records. The address is usually used as a buffer pointer in the
COBOL E35 exit routine. By using it in the E15 LINKAGE SECTION, MFX is
able to use the same parameter list for both COBOL exits E15 and E35.
• For the fifth definition specify a dummy area.

Syncsort MFX Programmer’s Guide 7–13

 • For the sixth definition (LEN-RU) specify data-name PIC 9(8)
COMPUTATIONAL. This is where MFX passes the length of the SORTIN
record to the COBOL exit.
• For the seventh definition (LEN-WK) specify data-name PIC 9(8)
COMPUTATIONAL. This is where the E15 routine passes the length of the
work area record to MFX.
• For the eighth definition define a dummy area.
• For the ninth definition (COMM-LEN) specify PIC 9(4) COMPUTATIONAL.
This area defines the communication area length.
• For the tenth definition (COMMUNICATION-AREA) code an OCCURS clause
DEPENDING ON data-name PIC X.

The IDENTIFICATION, ENVIRONMENT, and DATA Divisions

As always, the COBOL program must contain the entries required by the compiler
for these program divisions. Code the optional entries in these divisions according
to the requirements of the application.

The WORKING-STORAGE SECTION

If the exit routine inserts records into the final merge and replaces records passed
from MFX, the insertion record and the replacement record may be defined in this
section. These records will be moved to the WORK area described in the LINKAGE
SECTION, so be sure that the PICTURE clause or the OCCURS clause in the
WORK area is correct for these records.
This section may also define the return codes as 77-level data items. Alternatively,
these codes can be specified as literals in the MOVE instruction. (MOVE literal to
RETURN-CODE.) Note that RETURN-CODE is the name of a predefined storage
area in COBOL used to pass return codes to the sort; RETURN-CODE should not
be defined in the exit routine.

The PROCEDURE DIVISION

Specify the USING option on the PROCEDURE DIVISION header. Each identifier
specified after USING must be the same as those described in the 01-level of the
LINKAGE SECTION. Taking for example the identifiers defined in the fixed-
length record LINKAGE SECTION shown here, they would appear as:
PROCEDURE DIVISION USING EXIT-STATUS, RECORD-UP, WORK,
DUMMY1, DUMMY2, DUMMY3, DUMMY4, DUMMY5, COMM-LEN,
COMMUNICATION-AREA.
The GOBACK statement is used to return control to MFX. Do not use the EXIT
statement as it will cause unpredictable results. Be sure that MFX receives a valid
return code before the GOBACK statement is executed.

7–14 Syncsort MFX Programmer’s Guide

EXIT-STATUS Codes (Fixed and Variable-Length Records)
00 First record. MFX uses this Code to indicate the first call to the
COBOL exit and that the first record from SORTIN is in the
RECORD-UP area.
04 Most records. This is used for all calls except the first one when there
are records in the RECORD-UP area. After Code 00 has been issued,
Code 04 is passed to the exit until there is no record for the sort to
pass to the RECORD-UP area.
08 All records passed. This indicates that the last SORTIN record has
already been processed by the exit. Do not attempt to reference the
record again. No more records will be passed to the exit routine. Note
that if the SORTIN data set is empty, 08 will be passed every time
including the first time.

RETURN-CODE Codes (Fixed and Variable-Length Records)

0 Accept this record. This instructs MFX to accept the (unaltered)
record in the RECORD-UP area.
4 Delete this record. MFX will delete the current record in the
RECORD-UP area.
8 Do not return to this exit. This instructs MFX to close the exit for the
remainder of the sort application. This return code might be used at
SORTIN end-of-file (Exit Status Code 08) to indicate that extra
records will not be added at this point. If SORTIN is present, the
current input record and all subsequent records will be processed by
MFX.
12 Insert a record. This instructs MFX to add the record in the WORK
area to the input data set just ahead of the current record in the
RECORD-UP area. When MFX returns control to the E15, the same
record will be in the RECORD-UP area. The exit routine can then
add another record from the WORK area or process the current
record in RECORD-UP.
16 Terminate MFX. MFX will end its program and return to the calling
program or the Supervisor. MFX will issue a completion code of 16 to
indicate that the sort was unsuccessful.
20 Replace current record. MFX will replace the current record in the
RECORD-UP area with the record in the WORK area. Be sure that
the record in the WORK area is valid before passing it to MFX.

Syncsort MFX Programmer’s Guide 7–15

To Change a Record
In order to change the record in the RECORD-UP area, first move it to the WORK
area. All changes are made to the WORK area copy, which replaces the record in
RECORD-UP when 20 is moved to RETURN-CODE.

Sample COBOL E15, Fixed-Length Records

IDENTIFICATION DIVISION.
PROGRAM-ID. E15FL13C.

ENVIRONMENT DIVISION.

CONFIGURATION SECTION.

SOURCE-COMPUTER. IBM-370.
OBJECT-COMPUTER. IBM-370.

INPUT-OUTPUT SECTION.
FILE-CONTROL.
I-O-CONTROL.
DATA DIVISION.
FILE SECTION.

WORKING-STORAGE SECTION.
01 EVEN-FLAG PIC 9(2) VALUE ZERO.

01 USER-RETURN-CODE PIC 9(8) COMPUTATIONAL.

88 ACCEPT-REC VALUE 0.
88 DELETE-REC VALUE 4.
88 END-EXIT VALUE 8.
88 INSERT-REC VALUE 12.
88 END-SORT VALUE 16.
88 REPL-REC VALUE 20.

01 CHANGE-REC.
05 C-REC PIC X(6) VALUE 'CHANGE'.
05 C-INCR PIC 9(4) VALUE ZERO.
05 C-BLANK PIC X(90) VALUE ZERO.

01 INSRT-REC.
05 I-REC PIC X(6) VALUE 'INSERT'.
05 I-INCR PIC 9(4) VALUE ZERO.
05 I-BLANK PIC X(90) VALUE SPACES.

01 TOTAL.
05 BLANKS PIC X(10) VALUE ' E15'.
05 TITL PIC X(25) VALUE 'TOTAL RECORDS OUT'.
05 COUNTER PIC 9(8) VALUE 0.

Figure 344. (Page 1 of 2)Sample COBOL E15, Fixed-Length Record

7–16 Syncsort MFX Programmer’s Guide

 LINKAGE SECTION.
01 EXIT-STATUS PIC 9(8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 LAST-TIME VALUE 08.

01 RECORD-UP.
07 FILLER PIC 9(6).
07 R-SEQ1 PIC 9(2).
07 FILLER PIC X(92).
01 WORK PIC X(100).

PROCEDURE DIVISION USING EXIT-STATUS, RECORD-UP,WORK.

IF COUNTER GREATER THAN 100

MOVE 0 TO RETURN-CODE
GO TO RETURN-TO-SORT.
IF LAST-TIME GO TO RETURN-TO-SORT.
IF R-SEQ1 EQUAL 0
MOVE 4 TO RETURN-CODE
GO TO RETURN-TO-SORT.
IF R-SEQ1 EQUAL 5
MOVE 12 TO RETURN-CODE
MOVE INSRT-REC TO WORK
GO TO RETURN-TO-SORT.

IF R-SEQ1 EQUAL 6
MOVE 20 TO RETURN-CODE
MOVE CHANGE-REC TO WORK
GO TO RETURN-TO-SORT.
MOVE 0 TO RETURN-CODE.

RETURN-TO-SORT.
ADD 1 TO COUNTER.

IF LAST-TIME MOVE 8 TO RETURN-CODE

DISPLAY TOTAL.
GOBACK.

Figure 344. (Page 2 of 2)Sample COBOL E15, Fixed-Length Record

Syncsort MFX Programmer’s Guide 7–17

Sample COBOL E15, Variable-Length Records

IDENTIFICATION DIVISION.
PROGRAM-ID. E15VL19C.

ENVIRONMENT DIVISION.

CONFIGURATION SECTION.

SOURCE-COMPUTER. IBM-390.
OBJECT-COMPUTER. IBM-390.

INPUT-OUTPUT SECTION.
FILE-CONTROL.
I-O-CONTROL.

DATA DIVISION.
FILE SECTION.

WORKING-STORAGE SECTION.
01 EVEN-FLAG PIC 9(2) VALUE ZERO.

01 USER-RETURN-CODE PIC 9(8) COMPUTATIONAL.

88 ACCEPT-REC VALUE 0.
88 DELETE-REC VALUE 4.
88 END-EXIT VALUE 8.
88 INSERT-REC VALUE 12.
88 END-SORT VALUE 16.
88 REPL-REC VALUE 20.

01 CHANGE-REC.
05 C-REC PIC X(6) VALUE 'CHANGE'.
05 C-INCR PIC 9(4) VALUE ZERO.
05 C-BLANK PIC X(90) VALUE SPACES.

01 INSRT-REC.
05 I-REC PIC X(6) VALUE 'INSERT'.
05 I-INCR PIC 9(4) VALUE ZERO.
05 I-BLANK PIC X(90) VALUE SPACES.

01 TOTAL.
05 BLANKS PIC X(10) VALUE ' E15'.

Figure 345. (Page 1 of 2)Sample COBOL E15, Variable-Length Records

7–18 Syncsort MFX Programmer’s Guide

 05 TITL PIC X(25) VALUE 'TOTAL RECORDS OUT'.
05 COUNTER PIC 9(8) VALUE 0.

LINKAGE SECTION.
01 EXIT-STATUS PIC 9(8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.

01 RECORD-UP.
05 RU OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-RU PIC X.
01 WORK.
05 WK OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-WK PIC X.
01 IN-BUF PIC X(100).
01 DUMMY PIC 9(8) COMPUTATIONAL.
01 LEN-RU PIC 9(8) COMP.
01 LEN-WK PIC 9(8) COMP.

PROCEDURE DIVISION USING EXIT-STATUS, RECORD-UP, WORK,

IN-BUF, DUMMY, LEN-RU, LEN-WK.

IF NOT FIRST-TIME
ADD 1 TO COUNTER
MOVE 0 TO RETURN-CODE.

IF COUNTER LESS THAN 50

MOVE 54 TO LEN-WK
ADD 1 TO I-INCR
MOVE INSRT-REC TO WORK
MOVE 12 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 75

MOVE 44 TO LEN-WK
ADD 1 TO C-INCR
MOVE CHANGE-REC TO WORK
MOVE 20 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 100

MOVE 80 TO LEN-WK
ADD 1 TO I-INCR
MOVE 12 TO RETURN-CODE
MOVE INSRT-REC TO WORK.
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 200

MOVE 4 TO RETURN-CODE
GO TO RETURN-TO-SORT.

RETURN-TO-SORT.
IF LAST-TIME MOVE 8 TO RETURN-CODE
DISPLAY TOTAL.
GOBACK.

Figure 345. (Page 2 of 2)Sample COBOL E15, Variable-Length Records

Syncsort MFX Programmer’s Guide 7–19

Coding a C E15 Exit Routine
A C E15 exit program is indicated by the MODS control statement.
Like any other E15 exit routine, the C E15 exit routine is called each time a record
is brought into Phase 1 of a sort or Phase 3 of a copy. MFX and the C exit
communicate through arguments defined in the function header. For example,
records are passed to the C routine by the address presented in the second
argument in the function parameter list. No storage is reserved in the exit program
because the records exist elsewhere.
The C E15 exit routine can be written using either the C370 V2R1 compiler with
the V2R2 C370 Library, the SAA AD/Cycle C370 V1R2 Compiler and Library or
using the C/C++ for MVS/ESA V3R1.1 or higher Compiler and Library. When using
the LE/370 run-time library modules, it may be necessary to account for this
additional storage by adjusting the b value of the Exit-Name parameter on the
MODS statement.

Exit Communication
The parameter list structure required for passing fixed-length and variable-length
records between the sort and the exit is detailed in the following section. The
parameter names used in the examples are arbitrary but each definition is
required. Complete sample programs showing the use of the argument lists are
presented following the discussion of the exit interface.

Fixed-Length Records - Function Definition

int E15exit ( int* exit_status,
struct_ru* record_up,
struct_ins_rep* work,
int* dummy1, int* dummy2, int* dummy3,
int* dummy4, int* dummy5,
int* comm_len,
struct_ca* communication_area)

Figure 346. Sample Fixed-Length Records - Function Definition

The following describes the parameters used in the preceding definition.

exit_status This parameter points to a variable containing one of
the following exit status codes:

00 First record. MFX uses this code to indicate the

first call to the C exit and that the first record from

7–20 Syncsort MFX Programmer’s Guide

 SORTIN is in the record_up area. If the SORTIN is
empty or does not exist, a 08 status will be passed
the first time.

04 Most records. This is used for all calls except the

first one when there are records in the record_up
area. After Code 00 has been issued, Code 04 is
passed to the exit until there is no record for the
sort to pass to the record_up area.

08 All records passed. This indicates that the last

SORTIN record has already been processed by the
exit. Do not attempt to reference the record again.
No more records will be passed to the exit routine.
Note that if the SORTIN data set is empty or does
not exist, 08 will be passed every time including
the first time.
record_up The record_up parameter contains a pointer to the
record being passed to the E15 from the SORTIN. The
struct_ru data type represents a structure that
describes the fields within the SORTIN record.
work The work parameter contains a pointer to a work area
that is to be used to hold an inserted or replaced record
returned from the E15. The struct_ins_rep data type
represents a structure that describes the fields within
the inserted or replaced record.
dummy1 - dummy5 These parameters define unused place holders. They
are used with variable-length E15 and E35 communi-
cation. Their definition here allows a common parame-
ter list for fixed-length and variable-length C E15 and
E35 exits.
comm_len This parameter points to a variable that defines the
communication area length.
communication_area The communication_area parameter contains a
pointer to the communication area. The struct_ca data
type represents a structure that describes the fields in
the communication area.

Syncsort MFX Programmer’s Guide 7–21

Variable-Length Records - Function Definition
int E15exit ( int* exit_status,
void* record_up,
void* work,
int* dummy1, int* dummy2,
int* len_ru,
int* len_wk,
int* dummy3,
int* comm_len,
struct_ca* communication_area)

Figure 347. Sample Variable-Length Records - Function Definition

The following describes the parameters used in the preceding definition.

exit_status This parameter points to a variable containing exit
status codes. See the exit_status definition for a fixed-
length C E15 exit for the code definitions.
record_up The record_up parameter contains a “universal”
pointer to the record being passed to the E15 from the
SORTIN. The void* pointer can be cast to point an
appropriate structure to describe the record passed to
the exit. This allows different record structures, as is
common with variable-length records, to share a single
pointer definition.
work The work parameter contains a "universal" pointer to
a work area that is to be used to hold an inserted or
replaced record returned from the E15. The void*
pointer can be cast to point an appropriate structure to
describe the work record.
dummy1 - dummy3 These parameters define unused place holders. They
are used with C E35 communication. Their definition
here allows a common parameter list for C E15 and
E35 exits.
len_ru This parameter points to a variable that defines the
length of the SORTIN record passed to the E15. This is
the length of the record referred to in the record_up
parameter.
len_wk This parameter points to a variable that defines the
length of the record to be inserted or used as a replace-
ment for the record_up record. This is the length of the
record referred to in the work parameter. This field

7–22 Syncsort MFX Programmer’s Guide

 must be set by the exit when an insert or replace oper-
ation is performed.
comm_len This parameter points to a variable that defines the
communication area length.
communication_area The communication_area parameter contains a
pointer to the communication area. The struct_ca data
type represents a structure that describes the fields in
the communication area.

RETURN-CODE Codes (Fixed and Variable-Length Records)

The RETURN statement is used to return control to MFX. It must indicate one of
the following return values to indicate the action to be taken by MFX.
0 Accept this record. This instructs MFX to accept the (unaltered)
record in the record_up area.
4 Delete this record. MFX will delete the current record in the
record_up area.
8 Do not return to this exit. This instructs MFX to close the exit for the
remainder of the sort application. This return code might be used at
SORTIN end-of-file (exit_status code 08) to indicate that extra
records will not be added at this point. If SORTIN is present, the
current input record and all subsequent records will be processed by
MFX.
12 Insert a record. This instructs MFX to add the record in the work
area to the input data set just ahead of the current record in the
record_up area. When MFX returns control to the E15, the same
record will be in the record_up area. The exit routine can then add
another record from the work area or process the current record in
record_up. When inserting a variable-length record, insure that its
length is indicated in the len_wk parameter.
16 Terminate MFX. MFX will end its program and return to the calling
program or the Supervisor. MFX will issue a completion code of 16 to
indicate that the sort was unsuccessful.
20 Replace current record. MFX will replace the current record in the
record_up area with the record in the work area. Be sure that the
record in the work area is valid before passing it to MFX. When
replacing a variable-length record, insure that its length is indicated
in the len_wk parameter.

Syncsort MFX Programmer’s Guide 7–23

How to Change a Record
To change the record in the record_up area, first move it to the work area. All
changes are made to the work area copy, which replaces the record in record_up
when the return value from the exit is 20.

Sample C E15, Fixed-Length Records

#define FIRST_TIME 0
#define MOST_TIME 4
#define LAST_TIME 8
#define ACCEPT_REC 0
#define DELETE_REC 4
#define END_EXIT 8
#define INSERT_REC 12
#define END_SORT 16
#define REPL_REC 20
typedef _Packed struct record {
char name[6];
char code[4];
int serial_no;
} t_ru;
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
int SMPE15FB(int* exit_status,t_ru* record_up,t_ru* work,int* dummy1,
int* dummy2,int* dummy3,int* dummy4,int* dummy5,int* comm_len,
void* communication_area)
{
static counter=0;
int icode,return_code;
char * text1="CHANGE";
char * text2="INSERT";
if (counter > 10) {return_code=ACCEPT_REC;
goto return_to_sort;}
if (*exit_status == LAST_TIME) {return_code=END_EXIT;
goto return_to_sort;}

Figure 348. (Page 1 of 2)Sample C E15, Fixed-Length Record

7–24 Syncsort MFX Programmer’s Guide

 sscanf(record_up->code,"%4d",&icode);
if (icode==0) { return_code=DELETE_REC;
goto return_to_sort;}
if (icode==5) {
strncpy(work->name,text2,6);
sprintf(work->code,"%4d",icode+counter+8);
work->serial_no=300;
return_code=INSERT_REC;
goto return_to_sort;}
if (icode==6) {
strncpy(work->name,text1,6);
sprintf(work->code,"%4d",icode+1);
work->serial_no=record_up->serial_no+200;
return_code=REPL_REC;
goto return_to_sort;}
return_code=ACCEPT_REC;
return_to_sort:
counter++;
if (*exit_status==LAST_TIME)
{ return_code=END_EXIT;
printf("E15 total number of records handled:%d\n",counter);
}
return(return_code);
}

Figure 348. (Page 2 of 2)Sample C E15, Fixed-Length Record

Syncsort MFX Programmer’s Guide 7–25

Sample C E15, Variable-Length Records

#define FIRST_TIME 0
#define MOST_TIME 4
#define LAST_TIME 8
#define ACCEPT_REC 0
#define DELETE_REC 4
#define END_EXIT 8
#define INSERT_REC 12
#define END_SORT 16
#define REPL_REC 20
#define MAX_RLEN 104
typedef _Packed struct record1 {
char rec[6];
int incr;
char address[MAX_RLEN-14];
} t_ru1;
typedef _Packed struct record2 {
char title[10];
int number;
} t_ru2;
#include <stdio.h>
#include <stdlib.h>
#include <strings.h>
int SMPE15VB(int* exit_status,void* record_up,void* work,int* dummy1,
int* dummy2,int* len_ru,int* len_wk,int* dummy3,int* comm_len,
void* communication_area)
{
static counter=0,i_incr=0,i_number=0;
int return_code;
char *text1="CHANGE E15";
char *text2="INSERT E15";
t_ru1 * p_record1,*pwork1;
t_ru2 * p_record2,*pwork2;
p_record1 = (t_ru1 *)record_up;
pwork1 = (t_ru1 *)work;
p_record2 = (t_ru2 *)record_up;
pwork2 = (t_ru2 *)work;
if (*exit_status != FIRST_TIME) {counter++;
return_code=ACCEPT_REC;}

Figure 349. (Page 1 of 3)Sample C E15, Variable-Length Records

7–26 Syncsort MFX Programmer’s Guide

 if (counter<50) {
if (*len_ru == 14)
{
*len_wk = 14;
i_number++;
pwork2->number=i_number;
strncpy(pwork2->title,text2,10);
}
else {
*len_wk = 54;
i_incr++;
pwork1->incr=i_incr;
strncpy(pwork1->rec,text2,6);
}
return_code=INSERT_REC;
goto return_to_sort;}
if (counter<75) {
if (*len_ru == 14)
{
*len_wk = 14;
pwork2->number=p_record2->number+1;
strncpy(pwork2->title,text1,10);
}
else {
*len_wk = 54;
pwork1->incr=p_record1->incr+1;
strncpy(pwork1->rec,text1,6);
}
return_code=REPL_REC;
goto return_to_sort;}
if (counter<100) {
if (*len_ru == 14)
{
*len_wk = 14;
i_number++;
pwork2->number=i_number;
strncpy(pwork2->title,text2,10);
}

Figure 349. (Page 2 of 3)Sample C E15, Variable-Length Records

Syncsort MFX Programmer’s Guide 7–27

 else {
*len_wk = 80;
i_incr++;
pwork1->incr=i_incr;
strncpy(pwork1->rec,text2,6);
}
return_code=INSERT_REC;
goto return_to_sort;}
if (counter<200) {
return_code=DELETE_REC;
goto return_to_sort;}
return_to_sort:
if (*exit_status==LAST_TIME)
{ return_code=END_EXIT;
printf("E15 total number of records handled:%d\n",counter);
}
return(return_code);
}

Figure 349. (Page 3 of 3)Sample C E15, Variable-Length Records

Exit E25 - Deleting, Changing, and Summing Records

MFX gives control to exit E25 each time it is about to place a record in a Phase 2
output sequence, except for the first record of that sequence. Because all or part of
the input data set may skip this phase, it may be necessary to include an E35 to do
the job of the E25 during Phase 3. If it is possible to use the SUM or DUPKEYS
control statement in place of the exit, this is recommended.
These constraints apply to the coding of an E25 exit routine:
• The exit may not add records.
• The exit may not change sort control fields.
• The exit may not destroy the contents of the parameter list.
MFX will place the address of a 2-word parameter list in Register 1 each time it
passes control to the E25 routine. The first word, which is on a fullword boundary,
will contain the address of the record about to leave Phase 2. The second word will
contain the address of the record that has already passed into the output area.
Note that the first byte of each word contains zeros.

7–28 Syncsort MFX Programmer’s Guide

 PARAMETER LIST
Word 1: Address of record leaving Phase 2
Word 2: Address of record already in output area

Figure 350. Parameter List for E25

In order to change the record leaving Phase 2, the E25 exit program must first
move it to a work area. (The record in the output area may be changed, but must be
left where it is.) To sum two records, place the sum in the output area record and
delete the record leaving Phase 2.
After the record pair has been processed by the E25, a return code is placed into
Register 15 and control returns to MFX.

Return Codes
0 Accept this record. To instruct MFX to accept the record leaving
Phase 2, whether changed or unchanged, place return code 0 into
Register 15. The (work area) address of the record to be accepted
must be placed into Register 1.
4 Delete this record. This tells MFX to delete the record about to leave
Phase 2. It is not necessary to place the address of this record in Reg-
ister 1. The next time MFX returns control to the exit program, the
address of a new record will be in word 1 of the parameter list but
word 2 will be unchanged. (This permits further summing, for exam-
ple.)
16 Terminate MFX. MFX will end its program and return to the calling
program or the supervisor. MFX will give you a completion code of 16
to indicate that the sort was unsuccessful.

Exit E35 - Adding, Deleting, and Changing Records

When an output data set is available, you can elect to incorporate this exit to add,
delete or change records at the end of Phase 3. In the absence of an output data set,
this exit has full responsibility for output processing and, under normal conditions,
will delete every record passed by the sort.
E35 record processing has these constraints:
• If a record is to be changed, it should first be moved to a work area.
• The exit program may not destroy the contents of the parameter list.
• A user exit may not take checkpoints.

Syncsort MFX Programmer’s Guide 7–29

Coding the E35 Exit Routine for an Invoked Sort/Merge/Copy
When MFX is initiated from an ATTACH, LINK or XCTL macro, there are two
ways to include an E35 exit routine: (1) code the E35 exit routine in line with the
invoking program and specify the address of its entry point in the appropriate
entry of the calling program’s parameter list; or (2) define the separately compiled
routine in the MODS control statement. When the exit routine is coded in line with
the invoking program, it must handle all output processing; MFX will ignore a
SORTOUT DD statement and an OUTFIL control statement, if present.

The E35 Parameter List

This exit routine is given control each time MFX is about to place a record in the
output area after the final merge. In passing control to the E35 exit routine, MFX
places the address of a parameter list in Register 1. The parameter list contains
three entries.
For 32-bit parameter lists, each entry is one word. The first byte of each word
contains binary zeros. The first word contains the address of the record about to
leave Phase 3; after the last record has been passed, this word will contain zeros.
The second word contains the address of the record already in the output area;
when the first record is passed, this word will contain zeros.
The third word contains the user address constant. It contains either the last value
set in it by an E15 exit routine or, if not modified by an E15 exit routine, the initial
value from the user exit address constant provided in the invoking parameter list.
If the value was obtained from the 24-bit invoking parameter list, it is limited to 24
bits with the high-order byte set to X'00'.
If the user exit address constant was not provided or if MFX was JCL-invoked, it
will contain binary zeros. This word may be changed by the E35 exit routine
whenever it is entered, and it will remain the same on all subsequent entries to the
E35 exit routine.

PARAMETER LIST
1) Word 1: Address of record leaving Phase 3
2) Word 2: Address of record in output area
3) Word 3: User exit address constant.

Figure 351. Parameter List for E35

The 64-bit E35 parameter list can be enabled from the 64-bit sort invocation
parameter list. See p. 7.4 for information on returning records to MFX in register 1
and see ”Register Conventions” on p 7.4 for a description of the register 13 save

7–30 Syncsort MFX Programmer’s Guide

 area for exits called with the 64-bit exit parameter list. The same information is
passed to the exit as for the 32-bit parameter list, but double-word entries are used:

PARAMETER LIST
1) Word 1: X’00000000’ Word 2: Address of record leaving Phase 3
2) Word 3: X’00000000’ Word 4: Address of record in output area
3) Word 5: X’00000000’ Word 6: User exit address constant

Figure 352. 64-bit Parameter List for E35

Return Codes
0 Accept this record. This instructs MFX to accept the record now leav-
ing Phase 3. Place the (work area) address of this record in Register
1. Exits using the 32-bit parameter list return a 31-bit address; exits
using the 64-bit parameter list return a 64-bit address. This return
code is used when selectively editing records for output; it passes the
(possibly altered) record back to the sort. The RECORD statement is
required if this exit routine changes the maximum record length.
4 Delete this record. MFX will delete the record leaving Phase 3. There
is no need to load the address of this record into Register 1. When
MFX returns control to the E35, the first word of the parameter list
(the address of the record leaving Phase 3) will refer to a new record,
but the second word (the address of the output area record) will be
unchanged.
8 Disconnect E35. This instructs MFX to process any remaining
records without showing them to the E35 exit. Register 1 is ignored
for processing this return code. When this return code is used at end-
of file (signalled by a zero address in the first word of the parameter
list), it indicates that E35 is also finished and will not add additional
records. When used before end-of-file, it indicates that MFX should
process the "current" record passed to the E35, and any subsequent
records, as if there were no E35 present. Note that when MFX is not
creating any output files (SORTOUT or SORTOFxx) and E35 is the
only "output", MFX terminates immediately, since any subsequent
records will never be seen. Note that if an XSUM or XDUP data set
was being created, it will only contain records generated prior to the
return code of 8.
12 Insert a record. This tells MFX to add a record just before the record
is about to leave Phase 3. Load the address of the inserted record into
Register 1. Exits using the 32-bit parameter list return a 31-bit
address; exits using the 64-bit parameter list return a 64-bit address.
When MFX returns control to the E35 exit routine, the first word of
the parameter list (the address of the record leaving Phase 3) will be

Syncsort MFX Programmer’s Guide 7–31

 unchanged, but the second word (the address of the output area
record) will refer to the inserted record. The exit routine can then
add another record or process the current one.
16 Terminate MFX. This tells MFX to end its program and return to the
calling program or the supervisor. MFX uses a completion code of 16
to indicate that the sort was unsuccessful.

Coding a COBOL E35 Exit Routine

A COBOL E35 exit program can be indicated through the EXEC statement PARM
option (PARM='E35=COB'), the MODS control statement or the $ORTPARM DD
statement.
Like any other E35 exit routine, the COBOL E35 is called each time a record is
brought out of Phase 3. Communication between MFX and the COBOL exit takes
place in the LINKAGE SECTION of the COBOL program. For example, records are
passed to the COBOL routine in the second definition (RECORD-UP) area of the
LINKAGE SECTION. No storage is reserved in the exit program because the
records exist elsewhere.
If the COBOL exit routine uses any verb (EXHIBIT, DISPLAY, TRACE) which
results in output to the SYSOUT DD statement, there is a potential conflict with
MFX’s use of this DD statement. It is therefore recommended that you separate the
output by using either MFX’s MSGDD PARM option or the COBOL compiler’s
SYSx parm.

The LINKAGE SECTION

The LINKAGE SECTION examples that follow show the parameters required for
passing fixed-length and variable-length records to the sort. The data-names and
conditional names used in the examples are arbitrary but each definition is
required. The complete programs from which the examples are taken follow the
discussion of the exit.

7–32 Syncsort MFX Programmer’s Guide

Example 1: Fixed-Length Records

LINKAGE SECTION.
01 EXIT-STATUS PIC 9(8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.

01 RECORD-UP.
05 RU PIC X(100).

01 WORK.
05 WK PIC X(100).

01 IN-BUF.
05 IB PIC X(100).

01 DUMMY1 PIC X(4).

01 DUMMY2 PIC X.
01 DUMMY3 PIC X.
01 DUMMY4 PIC X.

01 COMM-LEN PIC 9(4) COMPUTATIONAL.

01 COMMUNICATION-AREA.

05 COMM-AREA OCCURS 1 TO 256 TIMES

DEPENDING ON COMM-LEN PIC X.

Figure 353. Sample Fixed-Length Records

The PICTURE and VALUE clauses for (1) the record passed from MFX, (2) the
record WORK area, and (3) the record in the output buffer are application-specific.
• For the first definition (EXIT-STATUS) specify PIC 9(8) COMPUTATIONAL.
When using 88 levels to define exit status codes, specify values 00, 04, and 08.
• For the second definition (RECORD-UP) define the record leaving Phase 3.
• For the third definition (WORK) define the record that MFX is to put in the
output data set. This is the "work" area.
• For the fourth definition (IN-BUF) define the record in the output data set.
• For the fifth definition define a dummy area with PIC X(4).
• For the sixth through the eighth definition define dummy areas.
• For the ninth definition (COMM-LEN) specify PIC 9(4) COMPUTATIONAL.
This area defines the communication area length.
• For the tenth definition (COMMUNICATION-AREA) code an OCCURS clause
DEPENDING ON data-name PIC X.

Syncsort MFX Programmer’s Guide 7–33

Example 2: Variable-Length Records

LINKAGE SECTION.
01 EXIT-STATUS PIC 9(8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.

01 RECORD-UP.
05 RU OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-RU PIC X.

01 WORK.
05 WK OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-WK PIC X.
01 IN-BUF.
05 IB OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-IB PIC X.

01 DUMMY PIC X(4)

01 LEN-RU PIC 9(8) COMPUTATIONAL.
01 LEN-WK PIC 9(8) COMPUTATIONAL.
01 LEN-IB PIC 9(8) COMPUTATIONAL.
01 COMM-LEN PIC 9(4) COMPUTATIONAL.

01 COMMUNICATION-AREA.
05 COMM-AREA OCCURS 1 TO 256 TIMES
DEPENDING ON COMM-LEN PIC X.

Figure 354. Sample Variable-Length Records

• For the first definition (EXIT-STATUS) specify PIC 9(8) COMPUTATIONAL.

When using 88 levels to define exit status codes, specify values 00, 04, and 08.
• For the second definition (RECORD-UP) code an OCCURS clause with the
DEPENDING ON data-name option specifying (1) the minimum and maximum
number of bytes of your variable-length records leaving Phase 3 (do not include
4 bytes for the RDW) and (2) DEPENDING ON data-name PIC X. Data-name is
defined in the sixth definition in the LINKAGE SECTION.
• For the third definition (WORK) code an OCCURS clause with the
DEPENDING ON data-name option specifying (1) the minimum and maximum
number of bytes for variable-length records you will pass to MFX (do not
include 4 bytes for the RDW) and (2) DEPENDING ON data-name PIC X. Data-
name is defined as the seventh definition in the LINKAGE SECTION. This area
is used for the “work” area.
• For the fourth definition (IN-BUF) define records in the output area. Code an
OCCURS clause with the DEPENDING ON data-name option specifying (1) the
minimum and maximum number of bytes for variable-length records in the
output data set (do not include 4-bytes for the RDW) and (2) DEPENDING ON

7–34 Syncsort MFX Programmer’s Guide

 data-name PIC X. Data-name is defined as the eighth definition in the
LINKAGE SECTION.
• For the fifth definition define a dummy area with PIC X(4).
• For the sixth definition (LEN-RU) specify PIC 9(8) COMPUTATIONAL. MFX
will pass the length of the record leaving Phase 3 in this area.
• For the seventh definition (LEN-WK) specify PIC 9(8) COMPUTATIONAL. The
E35 routine passes MFX the length of the record in the work area in this
section.
• For the eighth definition (LEN-IB) specify PIC 9(8) COMPUTATIONAL. MFX
passes the length of the record in the output area in this section.
• For the ninth definition (COMM-LEN) specify PIC 9(4) COMPUTATIONAL.
This area defines the communication area length.
• For the tenth definition (COMMUNICATION-AREA) code an OCCURS clause
DEPENDING ON data-name PIC X.

The IDENTIFICATION, ENVIRONMENT, and DATA Divisions

As always, the COBOL program must contain the entries required by the compiler
for these program divisions. Code the optional entries in these divisions according
to the requirements of the application.

The WORKING-STORAGE SECTION

If the exit routine inserts records into the final merge and replaces records passed
from MFX, the insertion record and the replacement record may be defined in this
section. These records will be moved to the WORK area described in the LINKAGE
SECTION, so be sure that the PICTURE clause or the OCCURS clause in the
WORK area is correct for these records.
This section may also define the return codes as 77-level data items. Alternatively,
these codes can be specified as literals in the MOVE instruction. (MOVE literal to
RETURN-CODE.) Note that RETURN-CODE is the name of a predefined storage
area in COBOL used to pass return codes to the sort; RETURN-CODE should not
be defined in the exit routine.

The PROCEDURE DIVISION

Specify the USING option on the PROCEDURE DIVISION header. Each identifier
specified after USING must be the same as those described in the 01-level of the
LINKAGE SECTION. Taking for example the identifiers defined in the fixed-
length record LINKAGE SECTION shown here, they would appear as:
PROCEDURE DIVISION USING EXIT-STATUS, RECORD-UP, WORK, IN-BUF,
DUMMY1, DUMMY2, DUMMY3, DUMMY4, COM-LEN, COMMUNICATION-
AREA.

Syncsort MFX Programmer’s Guide 7–35

 The GOBACK statement is used to return control to MFX. Do not use the EXIT
statement as it will cause unpredictable results. Be sure that MFX receives a valid
return code before the GOBACK statement is executed.

EXIT-STATUS Codes (Fixed and Variable-Length Records)

00 First Record. MFX uses this Code to indicate the first call to the
COBOL exit and that the first record to leave Phase 3 is in the
RECORD-UP area.
04 Most records. This is used for all calls except the first one when there
are records in the RECORD-UP area. After Code 00 has been issued,
Code 04 is passed to the exit until there is no record for the sort to
pass to the RECORD-UP area.
08 All records passed. This indicates that the last record has already
been processed by the exit. Do not attempt to reference the record
again. No more records will be passed to the exit routine. Note that if
MFX is not passing any records to Phase 3, 08 will be passed every
time including the first time.

RETURN-CODE Codes (Fixed and Variable-Length Records)

0 Accept this record. This instructs MFX to accept the (unaltered)
record in the RECORD-UP area.
4 Delete this record. MFX will delete the current record in the
RECORD-UP area.
8 Disconnect E35. This instructs MFX to process any remaining
records without showing them to the E35 exit. Register 1 is ignored
for processing this return code.
When this return code is used at end-of-file (signalled by EXIT-STA-
TUS LAST-TIME), it indicates that the E35 is also finished and will
not add additional records. When used before end-of-file, it indicates
that MFX should process the "current" record passed to the E35, and
any subsequent records, as if there were no E35 present. Note that
when MFX is not creating any output files (SORTOUT or SOR-
TOFxx) and E35 is the only "output," MFX terminates immediately,
since any subsequent records will never be seen. Also note that if an
XSUM or XDUP data set was being created, it will only contain
records generated prior to the return code of 8.
12 Insert a record. This instructs MFX to add the record in the WORK
area to the input data set just ahead of the current record in the
RECORD-UP area. When MFX returns control to the E35, the same
record will be in the RECORD-UP area. The exit routine can then
add another record from the WORK area or process the current
record in RECORD-UP.

7–36 Syncsort MFX Programmer’s Guide

 16 Terminate MFX. MFX will terminate and return to the calling pro-
gram or the Supervisor. MFX will issue a completion code of 16 to
indicate that the sort was unsuccessful.
20 Replace current record. MFX will replace the current record in the
RECORD-UP area with the record in the WORK area. Be sure that
the record in the WORK area is valid before passing it to MFX.

To Change a Record
In order to change the record in the RECORD-UP area, first move it to the WORK
area. Make the changes there and then pass return code 20 in RETURN-CODE.
The altered record in the WORK area will replace the record in RECORD-UP.

Syncsort MFX Programmer’s Guide 7–37

Sample COBOL E35, Fixed-Length Records

IDENTIFICATION DIVISION.

PROGRAM-ID. E35FL101.

ENVIRONMENT DIVISION.

CONFIGURATION SECTION.

SOURCE-COMPUTER. IBM-390.
OBJECT-COMPUTER. IBM-390.

INPUT-OUTPUT SECTION.
FILE-CONTROL.
I-O-CONTROL.

DATA DIVISION.
FILE SECTION.

WORKING-STORAGE SECTION.
01 EVEN-FLAG PIC 9(2) VALUE ZERO.

01 USER-RETURN-CODE PIC 9(8) COMPUTATIONAL.

88 ACCEPT-REC VALUE 0.
88 DELETE-REC VALUE 4.
88 END-EXIT VALUE 8.
88 INSERT-REC VALUE 12.
88 END-SORT VALUE 16.
88 REPL-REC VALUE 20.

01 CHANGE-REC.
05 C-REC PIC X(6) VALUE 'CHANGE'.
05 C-INCR PIC 9(4) VALUE ZERO.
05 C-BLANK PIC X(90) VALUE SPACES.

01 INSRT-REC.
05 I-REC PIC X(6) VALUE 'INSERT'.
05 I-INCR PIC 9(4) VALUE ZERO.
05 I-BLANK PIC X(90) VALUE SPACES.

01 TOTAL.
05 BLANKS PIC X(10) VALUE 'E35'.
05 TITL PIC X(25)
VALUE 'TOTAL RECORDS HANDLED'.
05 COUNTER PIC 9(8) VALUE 0

Figure 355. (Page 1 of 2)Sample COBOL E35, Fixed-Length Records

7–38 Syncsort MFX Programmer’s Guide

 LINKAGE SECTION.
01 EXIT-STATUS PIC 9(8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.

01 RECORD-UP.
05 RU PIC X(100).
01 WORK.
05 WK PIC X(100).
01 IN-BUF.
05 IB PIC X(100).
01 DUMMY1 PIC X(4).

PROCEDURE DIVISION USING EXIT-STATUS, RECORD-UP, WORK,

IN-BUF, DUMMY.

IF NOT FIRST-TIME
ADD 1 TO COUNTER
MOVE 0 TO RETURN-CODE.

IF COUNTER LESS THAN 50

ADD 1 TO I-INCR
MOVE INSRT-REC TO WORK
MOVE 12 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 75

ADD 1 TO C-INCR
MOVE CHANGE-REC TO WORK
MOVE 20 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 100

ADD 1 TO I-INCR
MOVE INSRT-REC TO WORK
MOVE 12 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 200

MOVE 4 TO RETURN-CODE
GO TO RETURN-TO-SORT.

RETURN-TO-SORT.
IF LAST-TIME MOVE 8 TO RETURN-CODE
DISPLAY TOTAL.
GOBACK.

Figure 355. (Page 2 of 2)Sample COBOL E35, Fixed-Length Records

Syncsort MFX Programmer’s Guide 7–39

Sample COBOL E35, Variable-Length Records

IDENTIFICATION DIVISION.

PROGRAM-ID. E35VL101.

ENVIRONMENT DIVISION.

CONFIGURATION SECTION.

SOURCE-COMPUTER. IBM-390.
OBJECT-COMPUTER. IBM-390.

INPUT-OUTPUT SECTION.
FILE-CONTROL.
I-O-CONTROL.

DATA DIVISION.
FILE SECTION.

WORKING-STORAGE SECTION.
01 EVEN-FLAG PIC 9(2) VALUE ZERO.

01 USER-RETURN-CODE PIC 9(8) COMPUTATIONAL.

88 ACCEPT-REC VALUE 0.
88 DELETE-REC VALUE 4.
88 END-EXIT VALUE 8.
88 INSERT-REC VALUE 12.
88 END-SORT VALUE 16.
88 REPL-REC VALUE 20.

01 CHANGE-REC.
05 C-REC PIC X(6) VALUE 'CHANGE'.
05 C-INCR PIC 9(4) VALUE ZERO.
05 C-BLANK PIC X(90) VALUE SPACES.

01 INSRT-REC.
05 I-REC PIC X(6) VALUE 'INSERT'.
05 I-INCR PIC X(4) VALUE ZERO.
05 I-BLANK PIC X(90) VALUE SPACES.

01 TOTAL.
05 BLANKS PIC X(10) VALUE ' E35'.
05 TITL PIC X(25)
VALUE 'TOTAL RECORDS HANDLED'.
05 COUNTER PIC 9(8) VALUE 0.

Figure 356. (Page 1 of 3)Sample COBOL E35, Variable-Length Records

7–40 Syncsort MFX Programmer’s Guide

 LINKAGE SECTION.
01 EXIT-STATUS PIC 9(8) COMPUTATIONAL.
88 FIRST-TIME VALUE 00.
88 MOST-TIME VALUE 04.
88 LAST-TIME VALUE 08.

01 RECORD-UP.
05 RU OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-RU PIC X.
01 WORK.
05 WK OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-WK PIC X.
01 IN-BUF.
05 IB OCCURS 1 TO 100 TIMES
DEPENDING ON LEN-IB PIC X.

01 DUMMY PIC X(4).

01 LEN-RU PIC 9(8) COMPUTATIONAL.
01 LEN-WK PIC 9(8) COMPUTATIONAL.
01 LEN-IB PIC 9(8) COMPUTATIONAL.

Figure 356. (Page 2 of 3)Sample COBOL E35, Variable-Length Records

Syncsort MFX Programmer’s Guide 7–41

 PROCEDURE DIVISION USING EXIT-STATUS, RECORD-UP, WORK,
IN-BUF, DUMMY, LEN-RU, LEN-WK, LEN-IB.

IF NOT FIRST-TIME
ADD 1 TO COUNTER
MOVE 0 TO RETURN-CODE.

IF COUNTER LESS THAN 50

MOVE 54 TO LEN-WK
ADD 1 TO I-INCR
MOVE INSRT-REC TO WORK
MOVE 12 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 75

MOVE 44 TO LEN-WK
ADD 1 TO C-INCR
MOVE CHANGE-REC TO WORK
MOVE 20 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 100

MOVE 80 TO LEN-WK
ADD 1 TO I-INCR
MOVE INSRT-REC TO WORK
MOVE 12 TO RETURN-CODE
GO TO RETURN-TO-SORT.

IF COUNTER LESS THAN 200

MOVE 4 TO RETURN-CODE
GO TO RETURN-TO-SORT.

RETURN-TO-SORT.
IF LAST-TIME MOVE 8 TO RETURN-CODE
DISPLAY TOTAL.
GOBACK.

Figure 356. (Page 3 of 3)Sample COBOL E35, Variable-Length Records

7–42 Syncsort MFX Programmer’s Guide

Coding a C E35 Exit Routine
A C E35 exit program is indicated by the MODS control statement.
Like any other E35 exit routine, the C E35 exit routine is called each time a record
is brought out of Phase 3. Communication between MFX and the C exit takes place
through arguments defined in the function header. For example, records are passed
to the C routine by an address presented in the second argument in the function
parameter list. No storage is reserved in the exit program because the records exist
elsewhere.
The C E35 exit routine can be written using either the C370 V2R1 compiler with
the V2R2 C370 Library, the SAA AD/Cycle C370 V1R2 Compiler and Library or the
C/C++ for MVS/ESA V3R1.1 Compiler and Library. When using the LE/370 run-
time library modules, it may be necessary to account for this additional storage by
adjusting the b value of the Exit-Name parameter on the MODS statement.

Exit Communication
The parameter list structure required for passing fixed-length and variable-length
records between the sort and the exit is detailed in the following section. The
parameter names used in the examples are arbitrary but each definition is
required. Complete sample programs showing the use of the argument lists are
presented following the discussion of the exit interface.

Fixed-Length Records - Function Definition

int E35exit ( int* exit_status,
struct_ru* record_up,
struct_ins_rep* work,
struct_in_buf* in_buf,
int* dummy1, int* dummy2, int* dummy3, int* dummy4,
int* comm_len,
struct_ca* communication_area)

Figure 357. Fixed-Length Records - Function Definition

The following describes the parameters used in the preceding definition.

exit_status This parameter points to a variable containing one of
the following exit status codes:

00 First record. MFX uses this Code to indicate the

first call to the C exit and that the first record to

Syncsort MFX Programmer’s Guide 7–43

 leave Phase 3 is in the record_up area. If there are
no records to pass to the exit, a 08 status will be
passed to the exit on the first call.

04 Most records. This is used for all calls except the

first one when there are records in the record_up
area. After Code 00 has been issued, Code 04 is
passed to the exit until there is no record for the
sort to pass to the record_up area.

08 All records passed. This indicates that the last

record has already been processed by the exit. Do
not attempt to reference the record again. No more
records will be passed to the exit routine. Note that
if MFX is not passing any records to Phase 3, 08
will be passed every time including the first time.
record_up The record_up parameter contains a pointer to the
record leaving Phase 3. The struct_ru data type rep-
resents a structure that describes the fields within the
record.
work The work parameter contains a pointer to a work area
that is to be used to hold an inserted or replaced record
returned from the E35. The struct_ins_rep data type
represents a structure that describes the fields within
the inserted or replaced record.
in_buf The in_buf parameter contains a pointer to the record
that MFX is to put in the output data set. Until a
record has been accepted or inserted, this pointer will
be null. A record at this address can be modified if
required.
dummy1 - dummy4 These parameters define unused place holders. They
are used with variable-length C E35 communication.
Their definition here allows a common parameter list
for fixed and variable-length C E15 and E35 exits.
comm_len This parameter points to a variable that defines the
communication area length.
communication_area The communication_area parameter contains a
pointer to the communication area. The struct_ca data
type represents a structure that describes the fields in
the communication area.

7–44 Syncsort MFX Programmer’s Guide

Variable-Length Records - Function Definition
int E35exit ( int* exit_status,
void* record_up,
void* work,
void* in_buf,
int* dummy1,
int* len_ru,
int* len_wk,
int* len_ib,
int* comm_len,
struct_ca* communication_area)

Figure 358. Variable-Length Records - Function Definition

The following describes the parameters used in the preceding definition.

exit_status This parameter points to a variable containing exit
status codes. See the exit_status definition for a fixed-
length C E35 exit for the code definitions.
record_up The record_up parameter contains a “universal”
pointer to the record leaving Phase 3. The void*
pointer can be cast to point an appropriate structure to
describe the record passed to the exit. This allows
different record structures, as is common with
variable-length records, to share a universal pointer.
work The work parameter contains a "universal" pointer to
a work area that is to be used to hold an inserted or
replaced record returned from the E35. The void*
pointer can be cast to point an appropriate structure to
describe the work record.
in_buf The in_buf parameter contains a "universal" pointer to
the record that MFX is to put in the output data set.
Until a record has been accepted or inserted, this
pointer will be null. The void* pointer can be cast to
point an appropriate structure to describe the work
record.
dummy1 This parameter defines an unused place holder.
len_ru This parameter points to a variable that defines the
length of the record leaving Phase 3. This is the length
of the record referred to in the record_up parameter.
len_wk This parameter points to a variable that defines the
length of the record to be inserted or used as a replace-

Syncsort MFX Programmer’s Guide 7–45

 ment for the record_up record. This is the length of the
record referred to in the work parameter.
len_ib This parameter points to a variable that defines the
length of the record that MFX is to put in the output
data set. This is the length of the record referred to in
the in_buf parameter.
comm_len This parameter points to a variable that defines the
communication area length.
communication_area The communication_area parameter contains a
pointer to the communication area. The struct_ca data
type represents a structure that describes the fields in
the communication area.

RETURN-CODE Codes (Fixed and Variable-Length Records)

0 Accept this record. This instructs MFX to accept the (unaltered)
record in the record_up area.
4 Delete this record. MFX will delete the current record in the
record_up area.
8 Disconnect E35. This instructs MFX to process any remaining
records without showing them to the E35 exit. When this return code
is used at end-of-file (signalled by exit_status 08), it indicates that
the E35 is also finished and will not add additional records. When
used before end-of-file, it indicates that MFX should process the "cur-
rent" record passed to the E35, and any subsequent records, as if
there were no E35 present. Note that when MFX is not creating any
output files (SORTOUT or SORTOFxx) and E35 is the only "output,"
MFX terminates immediately, since any subsequent records will
never be seen.
12 Insert a record. This instructs MFX to add the record in the work
area to the input data set just ahead of the current record in the
record_up area. When MFX returns control to the E35, the same
record will be in the record_up area. The exit routine can then add
another record from the work area or process the current record in
record_up.
16 Terminate MFX. MFX will terminate and return to the calling pro-
gram or the Supervisor. MFX will issue a completion code of 16 to
indicate that the sort was unsuccessful.
20 Replace current record. MFX will replace the current record in the
record_up area with the record in the work area. Be sure that the
record in the work area is valid before passing it to MFX.

7–46 Syncsort MFX Programmer’s Guide

Change a Record
In order to change the record in the record_up area, first move it to the provided
work area. Make the changes there and then pass return code 20. The altered
record in the work area will replace the record in record_up.

Sample C E35, Fixed-Length Records

#define FIRST_TIME 0
#define MOST_TIME 4
#define LAST_TIME 8
#define ACCEPT_REC 0
#define DELETE_REC 4
#define END_EXIT 8
#define INSERT_REC 12
#define END_SORT 16
#define REPL_REC 20
#include <decimal.h>
typedef _Packed struct record {
char rec[6];
decimal(7,0) incr;
char address[90];
} t_ru;
int counter,i_incr;
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
int SMPE35FB(int* exit_status,t_ru* record_up,t_ru* work,t_ru* in_buf,
int* dummy1,int* dummy2,int* dummy3,int* dummy4,int* comm_len,
void* communication_area)
{
int return_code;
char *text1="CHANGE";
char *text2="INSERT";
if (*exit_status != FIRST_TIME) {counter++;
return_code=ACCEPT_REC;
}

Figure 359. (Page 1 of 2)Sample C E35, Fixed-Length Records

Syncsort MFX Programmer’s Guide 7–47

 if (counter<50) {
i_incr++;
work->incr=i_incr;
strncpy(work->rec,text2,6);
return_code=INSERT_REC;
goto return_to_sort;}
if (counter<75) {
work->incr=record_up->incr+1d;
strncpy(work->rec,text1,6);
return_code=REPL_REC;
goto return_to_sort;}
if (counter<100) {
i_incr++;
work->incr=i_incr;
strncpy(work->rec,text2,6);
return_code=INSERT_REC;
goto return_to_sort;}
if (counter<200) {
return_code=DELETE_REC;
goto return_to_sort;}
return_to_sort:
if (*exit_status==LAST_TIME)
{ return_code=END_EXIT;
printf("E35 total number of records handled:%d\n",counter);
}
return(return_code);
}

Figure 359. (Page 2 of 2)Sample C E35, Fixed-Length Records

7–48 Syncsort MFX Programmer’s Guide

Sample C E35, Variable-Length Records

#define FIRST_TIME 0
#define MOST_TIME 4
#define LAST_TIME 8
#define ACCEPT_REC 0
#define DELETE_REC 4
#define END_EXIT 8
#define INSERT_REC 12
#define END_SORT 16
#define REPL_REC 20
#define MAX_RLEN 104
typedef _Packed struct record1 {
char rec[6];
int incr;
char address[MAX_RLEN-14];
} t_ru1;
typedef _Packed struct record2 {
char title[10];
int number;
} t_ru2;
int counter,i_incr,i_number;
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
int SMPE35VB(int* exit_status,void* record_up,void* work,void* in_buf,
int* dummy,int* len_ru,int* len_wk,int* len_ib,int* comm_len,
void* communication_area)
{

Figure 360. (Page 1 of 3)Sample C E35, Variable-Length Records

Syncsort MFX Programmer’s Guide 7–49

 int return_code;
char *text1="CHANGE E35";
char *text2="INSERT E35";
t_ru1 * p_record1,*pwork1;
t_ru2 * p_record2,*pwork2;
p_record1 = (t_ru1 *)record_up;
pwork1=(t_ru1 *)work;
p_record2 = (t_ru2 *)record_up;
pwork2=(t_ru2 *)work;
if (* exit_status != FIRST_TIME) {counter++;
return_code=ACCEPT_REC;}
if (counter<50) {
if (*len_ru == 14)
{
*len_wk = 14;
i_number++;
pwork2->number=i_number;
strncpy(pwork2->title,text2,10);
} else
{
*len_wk = 54;
i_incr++;
pwork1->incr=i_incr;
strncpy(pwork1->rec,text2,6);
}
return_code=INSERT_REC;
goto return_to_sort;}
if (counter<75) {
if (*len_ru == 14)
{
*len_wk = 14;
pwork2->number=p_record2->number+1;
strncpy(pwork2->title,text1,10);
} else
{
*len_wk = 54;
pwork1->incr=p_record1->incr+1;
strncpy(pwork1->rec,text1,6);
}
return_code=REPL_REC;
goto return_to_sort;}

Figure 360. (Page 2 of 3)Sample C E35, Variable-Length Records

7–50 Syncsort MFX Programmer’s Guide

 if (counter<100) {
if (*len_ru == 14)
{
*len_wk = 14;
i_number++;
pwork2->number=i_number;
strncpy(pwork2->title,text2,10);
} else
{
*len_wk = 80;
i_incr++;
pwork1->incr=i_incr;
strncpy(pwork1->rec,text2,6);
}
return_code=INSERT_REC;
goto return_to_sort;}
if (counter<200) {
return_code=DELETE_REC;
goto return_to_sort;}
return_to_sort:

if (*exit_status==LAST_TIME)
{ return_code=END_EXIT;
printf("E35 total number of records handled:%d\n",counter);
}
return(return_code);

Figure 360. (Page 3 of 3)Sample C E35, Variable-Length Records

Exit E16-Taking Action on Insufficient Intermediate Storage

Exit E16 is given control in the event that the input data set is unable to fit into
intermediate storage. There is no parameter list. The E16 return code tells MFX
how to respond to the insufficient SORTWK problem.

Return Codes
0 Sort present records only. This instructs MFX to process only those
records presently contained on the intermediate storage devices. The
sort will receive message WER054I RCD IN xxxxxxxx, OUT
yyyyyyyy if the data is read from SORTIN directly, or WER055I
INSERT xxxxxxxx, DELETE yyyyyyyy if the data receives input exit
(E14 or E15) or INCLUDE/OMIT processing. The message’s RCD IN
or INSERT xxxxxxxx figure indicates how many records have been

Syncsort MFX Programmer’s Guide 7–51

 sorted. For a sort with no exits or INCLUDE/OMIT processing, the
remaining records may be sorted by running another job using the
SKIPREC=n parameter on the SORT control statement, skipping the
xxxxxxxx number of records. The new sort will start just where the
last one left off. The final output is obtained by running a MERGE
with the two SORTOUT data sets.
4 Try to sort all records. This tells MFX to continue to read in records
from the input data set. If there are very few records left, the sort
may complete successfully. If there are too many records to continue
the sort, MFX will terminate with a SORT CAPACITY EXCEEDED
message.
12 Terminate MFX. MFX will terminate immediately with a SORT
CAPACITY EXCEEDED message.

Exits E17, E27, and E37 - Closing Data Sets

These exits are unusual in that they are entered only once, at the end of their
associated phase. Because of this, they may be efficiently used to clean up after
other exit routines (e.g., to close data sets). There are no parameter lists or return
codes for these exits.

Exits E18, E38, and E39 - Checking Labels, Processing Read

or
Write Errors, End-of-File Routines, Special VSAM Process-
ing
These exits are mainly used for I/O error recovery routines. However, they may also
be used to check labels, to do end-of-file processing, and to provide various
information to the VSAM access method.

Exit E18 and E38 Programs

Exit E18 is only used for sorts and exit E38 only for merges or copies. Each exit is
entered exactly once, at the start of SORTIN processing. At this time, MFX checks
Register 1 for the address of a user parameter list specifying the various open and
error exit routines that you want MFX to include. MFX will then enter these
routines at the appropriate times during execution. Because use of these exits
forces the use of BSAM for the input file(s), performance may be adversely affected.

7–52 Syncsort MFX Programmer’s Guide

 The format of the parameter list is given below. More information on the DCB fields
can be found in the appropriate IBM publication.

Syncsort MFX Programmer’s Guide 7–53

 Byte 1 Byte 2 Byte 3 Byte 4

01 SYNAD field

02 EXLST field

03 00 00 EROPT code

04 EODAD field

00 00 00 00

Table 57. Parameter List for E18 and E38

The parameter list must begin on a fullword boundary and consist of an integral
number of words. With the exception of the required fullword of zeros used to
indicate the end of the parameter list, entries are optional. The first byte of each
word identifies the parameter:
SYNAD field Indicated by 01 in byte 1. The SYNAD field contains the address
of a synchronous read error routine, assembled as part of the exit
program. Note that you may not use Register 13 as a save area
pointer on entry to your routine. You must either provide your
own save area or use the SYNADAF macro instruction.
EXLST field Indicated by 02 in byte 1. The EXLST field contains the address
of a list of pointers to user routines that perform operations such
as label checking. Note that in the event that the list contains a
DCB-exit entry, it will not be entered during concatenated SOR-
TIN processing.
EROPT code Indicated by 03 in byte 1. Bytes 2 and 3 contain zeros, byte 4 the
EROPT code. This code tells MFX what action to take if it discov-
ers an uncorrectable read error on a non-VSAM input file.

X'00' Follow the EROPT code in the DCB parameter of

the DD statement that describes the data set con-
taining the error.

X '20' Terminate the program.

X'40' Skip the block containing the error.

X'80' Accept the block containing the error.

EODAD field Indicated by 04 in byte 1. The EODAD field contains the address
of an end-of-file routine. It can only be used with an E18 exit.

7–54 Syncsort MFX Programmer’s Guide

VSAM Input to E18 and E38
With VSAM input, these exits can be used to pass the addresses of various VSAM
exits or to insert passwords into VSAM input ACB’s. When control is returned to
the sort, Register 1 must contain the address of a parameter list:

X'05' 3 byte address of VSAM exit list

X'06' 3-byte address of password list

F'0' Fullword of zeros

Table 58. Sample VSAM Parameter List for E18 and E38
If both address entries are present, they may be in either order. Only one need be
present. (QSAM parameters will be ignored.) The password list referenced in the
parameter list is found in the exit routine and is formatted as follows:
2 bytes on a halfword boundary:

Number n of entries in list

Followed by n 16-byte entries:

8-byte DDname

8-byte Password

The exit routine must not alter this list. The sort may destroy the last byte of the
DD name field.
The exit list is built using the VSAM EXLST macro, which provides the addresses
of the VSAM exit routines. VSAM branches directly to the routines which must
return to VSAM via the address in Register 14.
To do EODAD processing with E38, write a LERAD exit and check for X'04' in the
FDBK field of the RPL: this indicates input EOD. This field is needed by the merge,
so it should not be altered when returning to VSAM.
The following example shows how to code the return to the sort.

Syncsort MFX Programmer’s Guide 7–55

 ENTRY E18
.
.
.
E18 LA 1,PARMLST
ST 1,24(13) R13 points to sort's save area
LM 14,12,12(13)
BR 14
CNOP 0,4
PARMLST DC X'01'
DC AL3(BSAMERR)
DC X'02'
DC AL3(EXLST)
DC X'03'
DC X'000080' EROPT code
DC X'04'
DC AL3 (BSAMEOD)
DC X'05'
DC AL3 (VSAMEXL)
DC X'06'
DC AL3 (PWDLST)
DC A(0)
.
.
.
VSAMEXL EXLST SYNAD=SYNAD,LERAD=LERAD
PWDLST DC H'2'
DC CL8'SORTIN' SORTIN ddname
DC CL8'INPASS' SORTIN password
DC CL8'SORTOUT' SORTOUT ddname
DC CL8'OUTPASS' SORTOUT password
SYNAD ... VSAM synch error rtn
LERAD ... VSAM logic error rtn
BSAMERR ... BSAM error rtn
EXLST ... EXLST address list
BSAMEOD ... BSAM end of data rtn

Figure 361. Sample E18 Program

Exit E39 Programs

Exit E39 is used mainly for SORTOUT write error routines. The exit is entered
once at the beginning of merge or copy processing or the start of sort Phase 3. At
this time MFX checks Register 1 for the address of a user parameter list specifying
the various routines that you want MFX to include. MFX will then enter these
routines at the appropriate times during execution. The use of an E39 exit forces
the use of BSAM on the output file; this may degrade performance somewhat.
The format of the parameter list is given below.

7–56 Syncsort MFX Programmer’s Guide

 Byte 1 Byte 2 Byte 3 Byte 4

01 SYNAD field

02 EXLST field

00 00 00 00

Table 59. Parameter List for E39

The parameter list must begin on a fullword boundary and consist of an integral
number of words. With the exception of the required fullword of zeros used to
indicate the end of the parameter list, entries are optional. The first byte of each
word identifies the parameter:
SYNAD field Indicated by 01 in byte 1. The SYNAD field contains the address of
a synchronous write error routine, assembled as part of the exit pro-
gram. Note that you may not use Register 13 as a save area pointer
on entry to your routine. You must either provide your own save area
or use the SYNADAF macro instruction.
EXLST field Indicated by 02 in byte 1. The EXLST field contains the address of a
list of pointers to user routines that perform operations such as label
checking. If the EXLST field is specified, CHECKPOINT processing
will not be performed by MFX.
Exit E39 may be used to supply a VSAM exit list or password list for the output file
in the same manner as described for exits E18 and E38. Note that unlike E18 there
is no EODAD field with this exit.

Exit E61 - Modifying the Collating Process

Exit 61 is used to alter the collating of all control fields specified as having an o
(order) value of E in the SORT/MERGE control statement. Note that an E61 exit
routine is called in Phase 1 for sort applications and in Phase 3 for merge
applications. Each time MFX encounters an order E control field, it moves a copy of
the control field to a work area and passes the copy’s address to the exit routine.
Thus, the E61 exit program processes a control field image while leaving the
original control field intact. An order E control field is collated in ascending order
according to its f (format) code and its E61 image. In order to code an effective E61
routine, you must be familiar with the standard data formats used by the operating
system.
For all order E control fields except BInary fields, the number of bytes in the
control field image will be the number specified as the l (length) value on the
SORT/MERGE control statement. BInary fields are left and right padded with
zeros to the nearest byte boundary. For example, a control field designated as

Syncsort MFX Programmer’s Guide 7–57

 5.3,1.4,BI,E receives three bits of padding on the left, one on the right, producing
an image 2 bytes long.
An E61 exit can process only the first 256 bytes of the control field image in a single
pass. If a control field image is more than 256 bytes long, the exit will be entered
more than once for that control field.
If AC is specified as the format of a control field on the SORT or MERGE
statement, MFX will translate the field to ASCII before the E61 routine is given
control. In order to use an E61 routine to modify what would be an AC control field,
specify the field as CH in the SORT or MERGE statement and translate the image
to ASCII after it is altered by the E61 exit routine.
There is no advantage to coding an E61 exit if the ALTSEQ control statement can
provide the needed collating modification. ALTSEQ changes the installation’s
alternate collating sequence, used for all control fields specified with the format
code AQ.
An E61 exit cannot be used with locale processing (LOCALE option enabled).

The Parameter List

Each time your routine is executed, MFX will place the address of a three-word
parameter list in Register 1. The parameter list will be on a fullword boundary. The
first word contains the number of the control field within the record in byte 4. The
second word contains the address of the control field in the work area in bytes 2, 3,
and 4. The third word contains the length of the control field in bytes 3 and 4. All
values are given in hexadecimal and the unused bytes are filled with zeros.

Byte 1 Byte 2 Byte 3 Byte 4

00 00 00 Number of control
field

00 Address of control field in work area

00 00 Length of control field

Table 60. Parameter List for E61

Lengthening a Control Field Image

The length of the control field image is completely determined by the length and
format code of the control field. Therefore, in order to provide a 12-byte PD image of
5 bytes from the original record, it is necessary for the SORT/MERGE control
statement to reference a 12-byte PD control field that contains the 5 desired bytes.
The extra 7 bytes are used to contain the "lengthened" image.

7–58 Syncsort MFX Programmer’s Guide

Shortening a Control Field Image
The length of the control field image is completely determined by the length and
format code of the control field. To shorten a control field image, specify the full
length of the original control field as the l (length) value in the SORT/MERGE
control statement. Then shorten each image by the same number of bytes and pad
it uniformly to the length of the original field. Be sure to pad each control field
image with the same leading or trailing character, and replace data in the control
field image with the same type of data as that in the actual control field.

Reversing a Collating Sequence

Every order E control field is collated according to its image and format code, in
ascending order. To collate the field in apparent descending order, complement the
control field image according to its format code before returning control to MFX.
For a BI or CH field, for example, complement the image with hexadecimal FF’s
before returning control to the sort.

Coding REXX Exits

The exit routines E15 and E35 can be coded in REXX.

REXX Variables Provided by MFX

MFX provides a number of special REXX variables to facilitate the development of
REXX exits. These variables offer a simple, efficient means of establishing
communication between the exit and the sort/merge.
To load these variables, the following command must be used when the exit is
called.
ADDRESS 'SYNCREXX' 'GIVE'
When the exit completes its work, the exit should use the following sequence of
commands to return the variables to MFX.
ADDRESS 'SYNCREXX' 'TAKE'
RETURN
The following table describes the special REXX variables.

Syncsort MFX Programmer’s Guide 7–59

 Variable Function

SYRECORD When the exit is entered, SYRECORD contains the current data
record. The exit can accept the record, modify it or add a new record;
SYACTION should be set accordingly.

If SYRECORD is null, then MFX has no data remaining. When this

happens, the exit can either CLOSE or continue to INSERT new
records.

SYACTION This variable must be set before the exit returns control to MFX. It
describes the disposition of the current record. Possible values for
SYACTION are as follows:
ACCEPT: Retain the current record with no modification.
REPLACE: Replace the current record with the contents of
the SYRECORD.
DELETE: Delete the current record.
INSERT: Insert the contents of the SYRECORD before the
current record.
CLOSE: Do not return to the exit.
ABEND: Terminate MFX.

If an E15 is providing all the input (SORTIN not present), the only
valid values for SYACTION are INSERT, CLOSE or ABEND.

SYEXITYP This variable will automatically be set to E15 or E35, depending on

which type of exit is being called.

SYGBLN1... These eight special variables are global variables. You can set these to
...SYGBLN8 any value provided that the value does not exceed 15 characters in
length. MFX will insure that these variables are preserved across calls
to the exit.

SYGBLSTR This is an additional global variable. You can set this to any value, pro-
vided the string does not exceed 1024 characters in length. MFX will
insure that this variable is preserved across calls to the exit.

Table 61. REXX Variables Provided by MFX

Sample REXX Exit

The following example illustrates a REXX exit that will count the number of
records that are passed to the exit:

7–60 Syncsort MFX Programmer’s Guide

 address 'SYNCREXX' 'GIVE'
if sygbln1='SYGBLN1' then sygbln1=0
if LENGTH(syrecord) > 0
then do
syaction='REPLACE'
sygbln1=sygbln1 + 1
end
else do
syaction='CLOSE'
say 'REXX' syexityp 'counted' sygbln1 'records'
end
address 'SYNCREXX' 'TAKE'
return

Figure 362. Sample REXX Exit Code

Syncsort MFX Programmer’s Guide 7–61

7–62 Syncsort MFX Programmer’s Guide
Chapter 8 The Flow of the Sort

This chapter briefly outlines the flow of control in the standard Disk Sort, incore
sort, merge, and copy. It describes the order in which MFX will process and act on
the PARMs, control statements and exit routines you have provided. Note that all
executions begin with Phase 0 processing and that a given MFX execution will skip
steps where appropriate (e.g., will skip a “Variable-length record sampling” step if
sorting fixed-length records or HISTOGRM length values are supplied). No
attempt has been made to indicate which steps are required of all Disk Sorts,
incore sorts, etc., or to indicate the nature or timing of any abend processing.

Phase 0

• Process PARMs, merging EXEC and $ORTPARM PARM specifications. The

EXEC statement/invoking program’s parameter list overrides the installation
defaults. $ORTPARM overrides the EXEC statement/invoking program’s
parameter list.
• Process control statements (from the $ORTPARM DD statement and either the
SYSIN DD statement or the invoking program’s parameter list).
• Link-edit user exits (if necessary).
• Validate SORTIN/SORTINnn, SORTJNF1/SORTJNF2, and SORTOUT/
SORTOFxx/SORTOFx/SORTXSUM/SORTXDUP DCB attributes.

• If COPY with JOINKEYS, GO TO ————————> JOIN

Non-Sorting
Phase 3

Syncsort MFX Programmer’s Guide 8–1

 • If MERGE or COPY, GO TO ——————————> Non-Sorting
non-JOIN
Phase 3

• variable-length record sampling: open SORTIN, do the sampling, close

SORTIN.

• GO TO ——————————————————> Phase 1

Phase 1
non-JOIN

• Load Phase 1 exits.

• Call E11.
• Call E18.
• Open SORTIN.
• Perform SKIPREC processing.
• On the record level:
• Read from SORTIN or DB2 database for DB2 query.
• Call E15.
• Perform INCLUDE/OMIT processing.
• Perform INREC processing.
• Perform STOPAFT processing.
• Call E61.
• Perform SUM or DUPKEYS processing.
• Call E14.
• Call E16.
• Close SORTIN.
• Call E17.

• GO TO——————————————————> INCORE SORT

Determination

8–2 Syncsort MFX Programmer’s Guide

 Phase 1
JOIN

• Open SORTJNF1, SORTJNF2.

• On the record level:
• Read from SORTJNF1, SORTJNF2.
• Perform JOINKEYS INCLUDE/OMIT parameter processing, SORTJNF1,
SORTJNF2.
• Perform JOIN processing.
• Perform INCLUDE/OMIT processing.
• Perform INREC processing.
• Perform STOPAFT processing.
• Perform SUM or DUPKEYS processing.
• Close SORTJNF1, SORTJNF2.

INCORE SORT
Determination

• If there is sufficient memory, GO TO ———————> Incore Sort

• Delete Phase 1 exits.

• If all strings can be merged at once, GO TO ————> Sorting

Phase 3

• GO TO ———————————————————> Phase 2

Incore Sort

• Delete Phase 1 exits.

• Load Phase 3 exits.
• Call E31.

Syncsort MFX Programmer’s Guide 8–3

 • Call E39.
• Open SORTOUT.
• On the record level:
• Call E35.
• Write to SORTOUT.
• Close SORTOUT.
• Call E37.
• Delete Phase 3 exits.

• GO TO ———————————————————> Program
Termination

Phase 2

• Load Phase 2 exits.

• Call E21.
• On the record level:
• Call E25.
• Perform SUM or DUPKEYS processing.
• Call E27.
• Delete Phase 2 exits.

• GO TO ———————————————————> Sorting
Phase 3

Sorting
Phase 3

• Load Phase 3 exits.

• Take checkpoint.
• Call E31.
• Call E39.
• Open SORTOUT, SORTOFxx, SORTOFx, SORTXDUP, and SORTXSUM.
• On the record level:

8–4 Syncsort MFX Programmer’s Guide

 • Perform SUM or DUPKEYS processing.
• Write to SORTXSUM or SORTXDUP.
• Perform OUTREC processing.
• Call E35.
• If SORTOUT, SORTOFxx or SORTOFx are present, then for each output
data set:
• Perform STARTREC/ENDREC processing.
• Perform SAMPLE processing.
• Perform INCLUDE/OMIT parameter processing.
• Perform SAVE processing.
• Perform SPLIT/SPLITBY/SPLIT1R processing.
• Perform SortWriter functions.
• Perform OUTREC processing.
• Perform ANSI control character processing.
• Write to SORTOUT, SORTOFxx, or SORTOFx.
• Call E37.
• Close all data sets.
• Delete Phase 3 exits.

• GO TO———————————————————> Program
Termination

Non-Sorting, non-
JOIN Phase 3

• Load user exits for the merge/copy.

• Call E31.
• Call E38.
• Call E39.
• Open all data sets.
• Perform SKIPREC processing (for a copy).
• On the record level:
• Read from SORTIN/SORTINnn or DB2 database for DB2 query or (for a
merge) call E32 for a record.

Syncsort MFX Programmer’s Guide 8–5

 • Call E15 (for a copy).
• Perform INCLUDE/OMIT processing.
• Perform INREC processing.
• Perform STOPAFT processing (for a copy).
• Call E61 (for a merge).
• Perform SUM or DUPKEYS processing (for a merge).
• Write to SORTXSUM or SORTXDUP (for a merge).
• Perform OUTREC processing.
• Call E35.
• If SORTOUT, SORTOFxx or SORTOFx are present, then for each output file:
• Perform STARTREC/ENDREC processing.
• Perform SAMPLE processing.
• Perform INCLUDE/OMIT parameter processing.
• Perform SAVE processing.
• Perform SPLIT/SPLITBY/SPLIT1R processing.
• Perform SortWriter functions.
• Perform OUTREC processing.
• Perform ANSI control character processing.
• Write to SORTOUT, SORTOFxx or SORTOFx.
• Call E37.
• Close all data sets.
• Delete user exits.

• GO TO ———————————————————> Program
Termination

JOIN Non-
Sorting Phase 3

• Open SORTJNF1, SORTJNF2.

• On the record level:
• Read from SORTJNF1, SORTJNF2.

8–6 Syncsort MFX Programmer’s Guide

 • Perform JOINKEYS INCLUDE/OMIT parameter processing, SORTJNF1,
SORTJNF2.
• Perform JOIN processing.
• Perform INCLUDE/OMIT processing.
• Perform INREC processing.
• Perform STOPAFT processing.
• Perform OUTREC processing.
• Call E35.
• If SORTOUT, SORTOFxx or SORTOFx are present, then for each output file:
• Perform STARTREC/ENDREC processing.
• Perform SAMPLE processing.
• Perform INCLUDE/OMIT parameter processing.
• Perform SAVE processing.
• Perform SPLIT/SPLITBY/SPLIT1R processing.
• Perform SortWriter functions.
• Perform OUTREC processing.
• Perform ANSI control character processing.
• Write to SORTOUT, SORTOFxx or SORTOFx.
• Close all data sets.
• Delete E35 exit if loaded.

• GO TO———————————————————> Program
Termination

Program
Termination

• Print MFX messages.

• END.

Syncsort MFX Programmer’s Guide 8–7

8–8 Syncsort MFX Programmer’s Guide
Chapter 9 MAXSORT

MAXSORT: A Maximum Capacity Sort

MAXSORT is a maximum capacity sort designed to sort amounts of data that are
too large for an ordinary sorting technique to process.
MAXSORT breaks up the sorting process into small, individual sorts. At the end of
each individual sort a natural breakpoint occurs. At this time, the sorted data is
written out on intermediate storage devices and it becomes possible to stop the
program without losing the results of the previous processing. At each breakpoint,
an operator may intervene to change program options.
When all the input to the sort has been read and has become individual sorted data
sets, this output becomes input to one or more merges. If all the data sets can be
merged at once, one final merge is performed. If all the data sets cannot be merged
at once, some of them will be combined in one or more intermediate merges. Then,
when all the data sets can be merged at one time, the final merge is performed and
the final sorted output is produced.
The diagrams on the following two pages illustrate the MAXSORT technique.

Syncsort MFX Programmer’s Guide 9–1

9–2 Syncsort MFX Programmer’s Guide
Syncsort MFX Programmer’s Guide 9–3
MAXSORT’s Advantages
• Without MAXSORT, overlarge sorts require tape work areas or force you to
segment the input and execute multiple disk sorts. With MAXSORT, any input
data set can be handled by one sort execution using disk work space.
• MAXSORT requires less disk space than ordinary sorts. Because MAXSORT
stores the output of each individual sort on tape, the same disk SORTWK files
can be used over and over again.
• Since the output of each individual sort is a completely sorted data set, the
original job may be interrupted for higher priority jobs without wasting
processing time.
• If a system or program failure occurs, whatever data sets have already been
produced are still usable. The job can be restarted at the last breakpoint, and
all previously produced data sets can be used without resorting.

Job Control Language

MAXSORT and Disk Sort have similar JCL requirements. To initiate MAXSORT
using job control statements, specify PARM='MAXSORT' on the EXEC statement.
A program-initiated sort requests MAXSORT by using a PARM card image in the
data set defined by the $ORTPARM DD statement. In either case, it may be
necessary to request additional main storage in order to use the MAXSORT
technique.

Sample EXEC Statement

//stepname EXEC PGM=SYNCSORT,PARM=MAXSORT

Figure 363. MAXSORT EXEC Statement

9–4 Syncsort MFX Programmer’s Guide

DD Statements
MAXSORT’s DD statement requirements are summarized in the following table.
As many as three additional types of DD statements may be needed. Note that
SORTWK files must be allocated only to disk devices.

//$ORTPARM DD Used to override PARM or control statement information.

//SYSIN DD Control statement data set. Required unless control statements

are supplied via an invoking program parameter list.

//SYSOUT DD Message data set. Required unless all messages are routed to
console.

//SORTWKnn DD Disk work area definition. Required unless DYNALLOC is spec-

ified or MAXSORT is restarted at a MERGE breakpoint.

//SORTIN DD SORT input data set. Required unless there is an E15. Ignored if
the invoking program supplies an inline E15 exit routine;
optional if the MODS statement activates an E15 exit routine.

//SORTOUT DD Output data set. Required unless there is an E35. Ignored if the
invoking program supplies an inline E35 exit routine; optional if
the MODS statement activates an E35 exit routine.

//SORTXDUP DD Output data set for records deleted by DUPKEYS. Required if

XDUP parameter used.

//SORTXSUM DD Output data set for records deleted by SUM. Required if XSUM
parameter used.

//SORTBKPT DD Breakpoint data set. Must be DASD. Required.

//SORTOU00 DD Required if intermediate output is on tape.

//SORTOUnn DD Required if intermediate output is on disk or if intermediate out-

put is on tape and DYNATAPE is not specified.

//SORTCKPT DD Checkpoint data set. Required if Checkpoint-Restart is used.

//SORTMODS DD Required if user exits are in SYSIN and if user exits are to be
//SYSLIN DD linkage-edited at execution time.
//SYSLMOD DD
//SYSPRINT DD

//ddname DD Required for exits unless the exit is inline in LINKLIB/JOBLIB/

STEPLIB or in SYSIN.

Table 62. MAXSORT DD Statements

When the RELEASE=ON parameter is active (either via specification or by

default) at the conclusion of the sort portion of a MAXSORT, most of the allocated
SORTWK space is free (however, in the case of invoked sorts or SORTWKs defined

Syncsort MFX Programmer’s Guide 9–5

 as OLD, the data set is returned to the size allocated at MAXSORT initiation
rather than the minimum size possible).
The SORTBKPT, SORTOU00, SORTOUnn and SORTCKPT DD statements are
discussed below. Refer to “Chapter 4 JCL and Sample JCL/Control Statement
Streams” for a discussion of the other DD statements, which are specified for
MAXSORT just as they would be specified for Disk Sort.

SORTBKPT DD Statement
The required SORTBKPT statement defines the breakpoint data set on which the
sort control information is stored. At the end of each individual sort or merge, a
breakpoint is reached and the information in the breakpoint data set is
automatically amended. However, when MAXSORT is program-invoked or when
any exit other than an E35 exit is being used, breakpoints cannot be taken.
Nevertheless, the SORTBKPT statement must be specified for all MAXSORTs,
even those for which a breakpoint/restart is not possible.

Allocating Disk Space for the Breakpoint Data Set

The breakpoint data set must be allocated on a direct access device. It is
recommended that space for the breakpoint data set be allocated in the job step
preceding the sort step. Or, the breakpoint data set may be pre-allocated in a
separate job.

Sample Allocation of the Breakpoint Data Set

The following example illustrates how disk space for a breakpoint data set might be
allocated as part of a MAXSORT job control stream. These two statements would
follow the JOB statement:

//ALLOC EXEC PGM=IEFBR14

//BKPTDATA DD DSN=BKPT.DATA,DISP=(NEW,CATLG),UNIT=SYSDA,
// SPACE=(1000,(100,50))

Figure 364. Sample Job Step Allocating Disk Space for a Breakpoint Data Set

In this example, the name of the breakpoint data set is BKPT.DATA. Because this
data set must be kept until MAXSORT has completed, DISP=(NEW,CATLG) has
been specified. Supplying approximately 100K bytes of primary space and allowing
for secondary allocation should be adequate.

9–6 Syncsort MFX Programmer’s Guide

Sample SORTBKPT DD Statement
The sample SORTBKPT DD statement which follows identifies the breakpoint data
set for which disk space has already been allocated.

//SORTBKPT DD DSN=BKPT.DATA,DISP=(OLD,KEEP)

Figure 365. Sample SORTBKPT DD Statement

This SORTBKPT DD statement defines the breakpoint data set for which disk
space has previously been allocated. The data set name must be the same name
which was specified when the space for the breakpoint data set was allocated.
DISP=(OLD,KEEP) is specified so that this statement does not have to be changed
if MAXSORT is restarted. The DCB information should not be coded because
MAXSORT supplies these values. If a VOLSER was coded when the disk space for
the breakpoint data set was allocated, the same VOLSER must be specified in the
SORTBKPT DD statement.

SORTOU00 DD Statement
The SORTOU00 DD statement is required for every MAXSORT application in
which the intermediate output is stored on tape.
The SORTOU00 DD statement defines the unit to be used for the output of the
individual sorts and intermediate merges. MAXSORT will create and name these
data sets which will eventually be merged to produce the final sorted output.
The data set names generated will have one of the two formats. If the TIMESTMP
option is not specified (the installation default), data set names will have the
format:

 Snn 
TDS.jobname.  
 Mnn 

Figure 366. Data Set Names Format for TIMESTMP Not Specified

If the TIMESMP option is specified at installation time, data set names will have
the format:

Syncsort MFX Programmer’s Guide 9–7

  Snn 
TDS.Ddddhhmm.jobname.  
 Mnn 

Figure 367. Data Set Names Format for TIMESTMP Specified

In either case, the S or M indicates whether the output is from the sort phase or
from the merge phase; nn is the relative number of the data set (01 to 99). The
Ddddhhmm time stamp refers to the time (Julian day, hour and minute) the sort
began. The prefix default (TDS.) can be changed by specifying the BKPTDSN
PARM option.
The following rules should be observed in coding the SORTOU00 DD statement:
• Specify DISP=(NEW,KEEP) and a permanent DSNAME and VOL=PRIVATE so
that a scratch tape is used and the volumes are unloaded. Failure to specify this
can result in the rewinding of the scratch tape and overwriting of the
intermediate sort output.
• Specify the DEFER option in the UNIT parameter so that mount messages to
the operator that do not pertain to MAXSORT are suppressed.
• SORTOU00 and SORTIN cannot share the same tape unit. However,
SORTOU00 and SORTOUT may share the same tape unit unless the
DYNATAPE PARM is specified. SORTIN and SORTOUT may always share the
same unit.
• If DYNATAPE is in effect, the tape unit name must be the same as the unit
name specified in the TAPENAME PARM.

Sample SORTOU00 DD Statement

//SORTOU00 DD DSN=PERM.OU00,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE

Figure 368. Sample SORTOU00 DD Statement

SORTOUnn DD Statements
The SORTOUnn DD statements allocate the tape units used as input to the merge
phase. They are required unless the DYNATAPE option (available only under z/OS)
is used.
Although it is not necessary to specify the SORTOUnn DD statements when
DYNATAPE is used, it is a good idea to pre-allocate at least two SORTOUnn data
sets when the DYNATAPE option is specified. This ensures that the minimum

9–8 Syncsort MFX Programmer’s Guide

 required number of tape units will be available for the merge phase. When
additional units are available, DYNATAPE will provide performance benefits.
The following rules should be observed in coding the SORTOUnn DD statements:
• At least two tape units must be allocated in the absence of DYNATAPE.
However, allocating more units will make the merge phase complete more
quickly.
• Each statement must be allocated to a unique tape drive. If DYNATAPE is in
effect, the tape unit name must be the same as the unit name specified in the
TAPENAME PARM.
• All the tape units must operate at the same recording density. For best
performance, multiple density units should be run at the highest density.
• For each SORTOUnn statement, replace the 'nn' with a two digit number
between 01 and 99. The numbers need not be consecutive.
• Specify DISP=(NEW,KEEP), a permanent DSNAME and VOL=PRIVATE so
that a scratch tape is used and the tape volumes are unloaded.
• Specify the DEFER option in the UNIT parameter so that mount messages to
the operator that do not pertain to MAXSORT are suppressed.

Sample SORTOUnn DD Statements

//SORTOU01 DD DSN=PERM.OU01,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU02 DD DSN=PERM.OU02,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE

Figure 369. Sample SORTOUnn DD Statements

Using Disk for Intermediate Output

In most cases, tape units will be used to store the intermediate output of the
individual sorts and intermediate merges. However, it may be desirable in some
circumstances to place intermediate output on disk. Assigning the intermediate
output to a mass storage subsystem will not compromise the sort’s efficiency.
Because these files will be written and read sequentially, paging will be minimal.
If disk is used for intermediate storage, the following rules should be observed:
• The SORTOU00 statement should not be coded.
• SORTOUnn data sets must be permanent data sets if breakpoint/restart is to
be attempted.

Syncsort MFX Programmer’s Guide 9–9

 • To determine how many SORTOUnn DD statements to supply, divide the total
number of bytes of sort input data by the number of bytes of SORTWK space
and add 2 to the result. This figure is the number of SORTOUnn statements to
supply. If too few SORTOUnn DD statements are supplied, MAXSORT will
terminate for restart at the point at which a new DD statement is needed.
• Each SORTOUnn DD statement must allocate enough primary and secondary
space to hold all of the data written during that intermediate sort.
• Extended format data sets are not supported for disk SORTOUnn data sets.
• Track overflow is not supported for disk SORTOUnn data sets. Record lengths
must not exceed the track capacity unless VS or VBS records are being
processed.

SORTCKPT DD Statement
This DD statement is required in order to restart a MAXSORT which is task-
invoked or includes a user exit routine because in these cases MAXSORT cannot be
restarted from a breakpoint. The standard OS/VS Checkpoint-Restart feature is
used. Both automatic Checkpoint-Restart and deferred Checkpoint-Restart
capabilities are supported (see “Chapter 14 Performance Considerations”).
Checkpoints are taken at the end of each intermediate sort or merge.
The SORTBKPT DD statement must be specified in addition to the SORTCKPT
DD statement even when MAXSORT cannot be restarted from a breakpoint.
A MAXSORT with an E35 exit does not require the SORTCKPT DD statement.

Control Statements
Control statements will only be accepted at the initial execution of MAXSORT. If
MAXSORT is restarted, the control statements cannot be changed. Except for
JOIN, JOINKEYS, MERGE and REFORMAT, all MFX control statements are
supported for MAXSORT. In addition, the RESTART subparameter of the
SEQNUM parameter and the IFTHEN WHEN=GROUP parameter are not
supported on an INREC statement.

9–10 Syncsort MFX Programmer’s Guide

PARM Options
The MAXSORT parameters described below may be specified on the EXEC
statement, the $ORTPARM DD statement, PARMTBLE or PARMEXIT, and may
be listed in any order.

BKPTDSN

 cc...c. 
BKPTDSN=  
 TDS. 

Figure 370. BKPTDSN Format

The BKPTDSN PARM is used to change the prefix of the data set names for the
output of the individual sorts and intermediate merges. TDS. is the delivered
default. The last character of the prefix must be a period. If the TIMESTMP option
was specified at installation time, up to 21 characters may precede that period.
Otherwise, up to 31 characters may be specified before the final period.

DYNATAPE

 DYNATAPE 
 NODYNATAPE 
 

Figure 371. DYNATAPE Format

DYNATAPE instructs MAXSORT to dynamically allocate any tapes needed as

input for the merge phase. DYNATAPE may be used instead of (or as a supplement
to) SORTOUnn DD statements.
NODYNATAPE, the default, disables dynamic allocation.

MAXSORT

MAXSORT

Figure 372. MAXSORT Format

This parameter is required in order to execute MAXSORT.

Syncsort MFX Programmer’s Guide 9–11

MAXWKSP

 MAX 
 
MAXWKSP=  nM 
 
n 

Figure 373. MAXWKSP Format

This option specifies the maximum amount of disk SORTWK space that MAXSORT
can use. When this parameter is used, MAXSORT will release excess space in order
to meet the figure you have specified.
When the RELEASE=ON parameter is active (either via specification or by
default) at the conclusion of the sort portion of a MAXSORT, most of the allocated
SORTWK space is freed (however, in the case of invoked sorts or SORTWKs defined
as OLD, the data set is returned to the size allocated at MAXSORT initiation
rather than the minimum size possible).
If MAX, the default value, is specified, all primary and secondary space which has
been allocated will be acquired. The MAXWKSP value may also be specified as a
decimal number of cylinders (n) or as a decimal number of megabytes (nM) of work
space.
If MAXWKSP is specified as n cylinders, MAXSORT will convert the specification
to an actual byte value. MAXSORT will multiply by n the capacity of a cylinder on
the disk allocated to the lowest-numbered SORTWKnn DD statement.
Note: MAXWKSP should be specified as greater than or equal to MINWKSP, if
specified.

MINWKSP

 500 
 
MINWKSP=  nM 
 
n 

Figure 374. MINWKSP Format

This option specifies the minimum amount of disk SORTWK space that MAXSORT
can use. If the MINWKSP value exceeds the primary allocation and sufficient
secondary allocation cannot be obtained to meet the MINWKSP value at the time of
execution, the sort terminates. It can be restarted later when more space is
available.
The MINWKSP value may be specified as a decimal number of cylinders (n) or a
decimal number of megabytes (nM) of work space.

9–12 Syncsort MFX Programmer’s Guide

 The default MINWKSP value is 500 cylinders.
If MINWKSP is specified as n cylinders, MAXSORT will convert the specification to
an actual byte value. MAXSORT will multiply by n the capacity of a cylinder on the
disk allocated to the lowest-numbered SORTWKnn DD statement.
Note: MINWKSP should be specified as less than or equal to MAXWKSP, if speci-
fied.

RESTART

 LAST 
 
RESTART=  NO 
 
 id 

Figure 375. RESTART Format

This parameter specifies the point at which restart is to occur.

LAST, the default value, requests that the sort start at the most recent breakpoint.
NO specifies that the SORTBKPT data set is to be cleared so that it can be used for
a new job. (Be sure to specify NO only when the SORTBKPT data set is empty or
should be destroyed.)
To restart at a particular breakpoint, code its id number. The breakpoint id number
is provided by message WER350I.

SORTSIZE

n 
 
SORTSIZE=  nM 
 
 nT 

Figure 376. SORTSIZE Format

This option is accepted but ignored. Its function has been replaced by MFX internal
techniques.

SORTTIME

n 
SORTTIME=  
 1440 

Figure 377. SORTTIME Format

Syncsort MFX Programmer’s Guide 9–13

 The SORTTIME parameter terminates the sort at the next breakpoint after n
minutes of clock time have elapsed. (The sort may be restarted later.) The default is
1440 minutes (24 hours).
If this parameter is omitted or 1440 is specified, the sort will not terminate
prematurely.
This parameter may be specified with operator communication at installation time.
If operator communication is specified, the sort will be interrupted at the next
breakpoint after the specified amount of time has elapsed and the operator will be
asked whether to terminate the sort or continue until the next breakpoint.

TAPENAME

 name 
TAPENAME=  
 TAPE 

Figure 378. TAPENAME Format

This parameter specifies the tape unit generic name for dynamic tape allocation.
The default TAPENAME is TAPE.
If the TAPENAME parameter is specified, the same unit generic name must be
specified for both the SORTOU00 and the SORTOUnn DD statements.
The tape unit generic name must be a valid unit name at your installation.

Exit Programs
All the exits available for Disk Sort are supported for MAXSORT. However, since
MAXSORT never runs out of work space on even the largest sorts, an E16 exit
routine will never be called. Exit routines may be written in COBOL, C, Assembler
language, or REXX. All exits should be prelink-edited for maximum efficiency.
The following rules must be observed when MAXSORT includes an exit routine:
• Exit programs are not allowed to take their own z/OS checkpoints.
• MAXSORT may take system checkpoints when the following exits are active:
E14, E15, E25, E35 and E61. Since a checkpoint may be taken between any two
calls of these exits, these routines should be coded accordingly. Any restrictions
that apply to system Checkpoint-Restart, such as restrictions on the use of data
sets, are applicable to the coding of these exit routines.

9–14 Syncsort MFX Programmer’s Guide

Invoking MAXSORT from a Program
MAXSORT can be invoked from programs written in COBOL, PL/1 or Assembler
language. However, this is the least efficient method of executing MAXSORT and
performance benefits will be realized if MAXSORT is initiated through job control
language.
When MAXSORT is invoked from a program, the MAXSORT PARM should be
specified in the $ORTPARM DD statement. The SYSIN DD statement is ignored.

Restarting MAXSORT
A JCL-initiated MAXSORT can be restarted from a breakpoint if necessary. When
MAXSORT is restarted from a breakpoint, the following PARM options cannot be
modified: CMP=CPD/CLC, EQUALS, E15/E35=COB, FILSZ, LOCALE, MAXSORT,
STOPAFT and TAPENAME. Other PARM options will be accepted if they are
specified on the EXEC statement. Only the CORE parameter can be passed
through $ORTPARM.
MFX control statements cannot be modified when MAXSORT is restarted.
However, the l5, l6 and l7 values on the LENGTH parameter of the RECORD control
statement can be altered.

Restarting MAXSORT with Exit Routines or an Invoked MAXSORT

When MAXSORT includes an exit routine or is invoked from a program, it cannot
be restarted from a breakpoint. Instead, it can be restarted from a checkpoint using
the standard OS/VS Checkpoint-Restart feature. Checkpoints are taken at the end
of each intermediate sort or merge.
When MAXSORT is restarted from a checkpoint, modified PARM options cannot be
specified on the EXEC statement. Only the CORE parameter can be passed
through $ORTPARM.
To specify that checkpoints be taken for a MAXSORT with an exit routine or for an
invoked MAXSORT, the following rules must be observed:
• Include the SORTCKPT DD statement in the JCL (in addition to the
SORTBKPT DD statement.
• Assign a permanent data set name to every SORTWKnn DD statement and
specify DISP=(NEW,DELETE,KEEP).
• Specify RD=R and MSGLEVEL=1 on the JOB statement.
• Specify the CKPT parameter on the SORT/MERGE control statement.

Syncsort MFX Programmer’s Guide 9–15

MAXSORT’s Operator Interface
If MAXSORT’s operator interface options are enabled when MFX is installed, they
will permit operator communication at selected breakpoints (e.g., at the first
breakpoint after SORTTIME has expired, or when tape drives are dynamically
allocated under DYNATAPE.) Operator communication allows the operator to
examine the environment at execution time to decide whether or not to terminate
MAXSORT at that breakpoint. If the operator decides to terminate the sort, it can
be restarted later at that breakpoint. All the previously produced sorted data sets
can be used without resorting.
Operator communication with MAXSORT is not a delivered default - these options
must be enabled at MFX installation time.
For example, if MAXSORT’s assigned block of computer time (its SORTTIME
value) has been exhausted and MFX was installed to permit operator intervention
at such times, message WER375D is generated.

WER375D PAYROLL.SORTSTEP - MAXSORT BKPT PAYROLL S12

WER375D TIME ESTIMATE: 30 MINUTES UNTIL NEXT NOTIFICATION
WER375D REPLY 'GO' TO CONTINUE, 'STOP' TO TERMINATE

Figure 379. Example: Operator Notification at SORTTIME Expiration

The operator receiving this message can decide to terminate the sort or allow it to
continue, basing his decision on scheduling priorities and the estimated time of the
sort. When another 30 minutes have passed, the operator will be asked again
whether or not MAXSORT should be terminated.
When DYNATAPE is specified and operator communication has been enabled at
installation time, message WER376D may be generated to report the results of the
dynamic allocation attempt.

WER376D PAYROLL.SORTSTEP - MAXSORT BKPT PAYROLL.S01

WER376D 4 TAPE UNITS ALLOCATED TO PAYROLL
WER376D 6 TAPE UNITS NEEDED FOR BEST PERFORMANCE
WER376D TIME ESTIMATE USING 4 TAPE UNITS--
WER376D 20 MINUTES TO NEXT BREAKPOINT
WER376D REPLY 'GO' TO CONTINUE, 'STOP' TO TERMINATE, 'NN' # UNITS

Figure 380. Example: Operator Notification with DYNATAPE

If the operator responds 'GO', MAXSORT will execute with four tape units. If the
operator responds 'STOP', MAXSORT will terminate. If the operator responds with
a number ('NN'), MAXSORT will try to allocate that total number of tape drives.
Ideally, the operator should specify six for 'NN' because MAXSORT needs six tape
units for best performance. If the operator requests additional tape units, message
WER376D will be reissued. The operator will again be prompted for a 'GO', 'STOP'

9–16 Syncsort MFX Programmer’s Guide

 or 'NN' reply. In this way, the operator can balance the requirements of MAXSORT
against the requirements of other jobs that are executing at the same time.
When DYNATAPE is specified, there may not be enough tape units available for
dynamic allocation. In this case, message WER377D is generated.

WER377D PAYROLL.SORTSTEP - MAXSORT BKPT PAYROLL.S01

WER377D INSUFFICIENT TAPE UNITS AVAILABLE
WER377D 2 TAPE UNITS ALLOCATED TO PAYROLL
WER377D 4 TAPE UNITS NEEDED TO CONTINUE EXECUTION
WER377D REPLY 'RETRY' TO GET UNITS, 'STOP' TO TERMINATE

Figure 381. Example: Operator Notification of Insufficient Tape Units under DYNATAPE

The operator receiving message WER377D can wait until additional tape drives
have been released and then reply 'RETRY'. Or, the operator can answer 'STOP' to
terminate the job and then restart it later when more tape drives become available.
If the DYNATAPE and TAPENAME PARMs have been specified and all tape units
on the system within the TAPENAME class have already been allocated, message
WER378D is generated.

WER378D NO ADDITIONAL TAPE UNITS EXIST FOR GENERIC CLASS 2400-3

Figure 382. Example: Operator Notification of Insufficient Tape Units for TAPENAME
Class

Message WER378D is followed by message WER376D if the number of tape drives

allocated is sufficient for execution, or by message WER377D if it is not sufficient.

Syncsort MFX Programmer’s Guide 9–17

Sample MAXSORT JCL/Control Streams
The following examples illustrate how the JCL could be coded for typical 100
gigabyte MAXSORTs.

9–18 Syncsort MFX Programmer’s Guide

Example 1: A 100 Gigabyte MAXSORT with only Minimal Disk Space
Available
An installation is running a 100 gigabyte sort and has a restricted amount of disk
space available for SORTWK across ten volumes (WORK1, WORK2, ...WORK10).
The JCL for this job follows.

//ALLOC EXEC PGM=IEFBR14 1

//BKPTDATA DD DSN=BKPT.DATA,DISP=(NEW,CATLG), 2
// UNIT=SYSDA,SPACE=(1000,(100,50))
//SORT EXEC PGM=SYNCSORT,PARM='MAXSORT,MINWKSP=6000' 3
//SORTBKPT DD DSN=BKPT.DATA,DISP=(OLD,KEEP) 4
//SYSOUT DD ... 5
//SORTIN DD ...
//SORTOUT DD ...
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(300,175)), 6
// VOL=SER=WORK1
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK2
//SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK3
//SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK4
//SORTWK05 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK5
//SORTWK06 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK6
//SORTWK07 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK7
//SORTWK08 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK8
//SORTWK09 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK9
//SORTWK10 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),
// VOL=SER=WORK10
//SORTOU00 DD DSN=PERM.OU00,DISP=(NEW,KEEP), 7
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU01 DD DSN=PERM.OU01,DISP=(NEW,KEEP), 8
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU02 DD DSN=PERM.OU02,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU03 DD DSN=PERM.OU03,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SYSIN DD * 9
SORT FIELDS=(1,10,CH,A)
/*

Figure 383. Sample JCL Control Stream for a 100 Gigabyte MAXSORT

Syncsort MFX Programmer’s Guide 9–19

 1. This job step is run in order to allocate the disk space for the breakpoint data set via
IBM utility program IEFBR14.

2. This statement allocates the space for the breakpoint data set. Specify (NEW,CATLG)
because this data set must be saved.

3. The EXEC statement initiates the regular MFX program, and the MAXSORT PARM is
specified, as required. This job requests a minimum of 6000 cylinders of disk space for
SORTWKnn data sets. If that much space cannot be obtained during the job, the
program will terminate.

4. The SORTBKPT DD statement is required for all MAXSORTs. It identifies the

breakpoint data set which was allocated in the first job step. DISP=(OLD,KEEP) is
specified so that this statement can be reused if MAXSORT is restarted.

5. These DD statements are coded just as they would be for an ordinary sort.

6. The SORTWKnn DD statements must be allocated to disk or MAXSORT will terminate.

In this case, 3000 cylinders of primary space have been allocated. Secondary allocation
could provide up to 2625 cylinders on each volume if that amount of free space exists.
Since the MINWKSP PARM specifies at least 6000 cylinders, this program will
terminate unless 3000 cylinders of secondary space can be obtained.

7. The SORTOU00 DD statement is required for this job because the intermediate sort
output will be stored on tape. DISP=(NEW,KEEP), a permanent DSN and
VOL=PRIVATE are specified to ensure that the system unloads each output tape. The
DEFER option in the UNIT parameter is specified so that mount messages to the
operator that do not pertain to MAXSORT are suppressed.

8. The SORTOU01, SORTOU02 and SORTOU03 DD statements allocate the tape units
used as input to the merge phase. Permanent DSNAMEs, DISP=(NEW,KEEP),
VOL=PRIVATE and the DEFER option in the UNIT parameter are all specified just as
they were for the SORTOU00 DD statement.

9. The sort control statements are included here.

9–20 Syncsort MFX Programmer’s Guide

Example 2: Restarting the MAXSORT in Example 1 from a Breakpoint
Example 1 can be restarted from a breakpoint simply by submitting the original job
control stream without the job step which allocated space for the breakpoint data
set. The job will be restarted from the last breakpoint because RESTART=LAST is
the default; it is not necessary to specify RESTART=LAST on the EXEC statement.
The JCL for this job follows.

//SORT EXEC PGM=SYNCSORT,PARM='MAXSORT,MINWKSP=6000'

//SORTBKPT DD DSN=BKPT.DATA,DISP=(OLD,KEEP)
//SYSOUT DD ...
//SORTIN DD ...
//SORTOUT DD ...
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK1
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK2
//SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK3
//SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK4
//SORTWK05 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK5
//SORTWK06 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK6
//SORTWK07 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK7
//SORTWK08 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK8
//SORTWK09 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK9
//SORTWK10 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK10
//SORTOU00 DD DSN=PERM.OU00,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU01 DD DSN=PERM.OU01,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU02 DD DSN=PERM.OU02,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU03 DD DSN=PERM.OU03,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SYSIN DD *
SORT FIELDS=(1,10,CH,A)
/*

Figure 384. Sample JCL Control Stream for Restarting a 100 Gigabyte MAXSORT

The JCL is identical to the JCL in Example 1 except that the step which allocated
the disk space for the breakpoint data set is not resubmitted.

Syncsort MFX Programmer’s Guide 9–21

Example 3: A 100 Gigabyte MAXSORT with Dynamic Tape Allocation
This example is identical to Example 1 with one difference: the DYNATAPE PARM
requests dynamic tape allocation.
The JCL for this job follows.

//ALLOC EXEC PGM=IEFBR14

//BKPTDATA DD DSN=BKPT.DATA,DISP=(NEW,CATLG),UNIT=SYSDA,
// SPACE=(1000,(100,50))
//SORT EXEC PGM=SYNCSORT,PARM='MAXSORT,MINWKSP=6000,
// DYNATAPE'
//SORTBKPT DD DSN=BKPT.DATA,DISP=(OLD,KEEP)
//SYSOUT DD ...
//SORTIN DD ...
//SORTOUT DD ...
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK1
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK2
//SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK3
//SORTWK04 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK4
//SORTWK05 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK5
//SORTWK06 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK6
//SORTWK07 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK7
//SORTWK08 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK8
//SORTWK09 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK9
//SORTWK10 DD UNIT=SYSDA,SPACE=(CYL,(300,175)),VOL=SER=WORK10
//SORTOU00 DD DSN=PERM.OU00,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU01 DD DSN=PERM.OU01,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SORTOU02 DD DSN=PERM.OU02,DISP=(NEW,KEEP),
// UNIT=(TAPE,,DEFER),VOL=PRIVATE
//SYSIN DD *
SORT FIELDS=(1,10,CH,A)
/*

Figure 385. Sample JCL Control Stream for a 100 Gigabyte MAXSORT

The DYNATAPE PARM requests that tape units be obtained dynamically. Because
DYNATAPE has been specified, the SORTOU01, SORTOU02, and SORTOU03 DD
statements specified in Example 1 do not have to be supplied. They will be created
and dynamically allocated when needed. If enough tape units are available at the
time the job is run, the sort will be successfully completed in one step.
However, there may not be enough tape devices available under dynamic allocation
at execution time. In that case, the job will terminate and can be restarted at a
later time when more tape units are available.
For best results, code two SORTOUnn DD statements in addition to specifying the
DYNATAPE PARM as the above example illustrates. This approach ensures that
MAXSORT will have the minimum two tape units needed for the merge phase and

9–22 Syncsort MFX Programmer’s Guide

 also allows MAXSORT to take advantage of the additional tapes available under
dynamic allocation.

Tuning MAXSORT
MAXSORT’s performance can be optimized by controlling the intermediate sorts
which it processes. A balance should be achieved between the number and duration
of intermediate sorts. Limiting the number of sorts reduces the required tape
mounts and restricting the duration of sorts decreases the interval between
breakpoints.
A good rule of thumb is that each intermediate sorted data set should create from
one to five volumes of input data, and the only way to determine the amount of
input data is by controlling the amount of SORTWK space used. This is illustrated
in Figure 386.

1 3590 tape volume can contain 20 gigabytes

1 3390 cylinder can hold approximately 800,000 bytes
SORTIN: 100 gigabytes (5 tape volumes)
OBJECTIVE: Each intermediate sort processes one input volume,
i.e. 5 intermediate sorts should be run.

To determine SORTWK allocation (for 3390 SORTWKs):

Divide one input volume 21,474,836,480 bytes
by cylinder capacity 800,000

Quotient: 26844 cylinders.

To ensure 5 intermediate sorts:

Allocate 26844 cylinders

Set MAXWKSP=21759M

Figure 386. Calculating MAXWKSP

If only 3,000 cylinders were allocated in the preceding example, 45 intermediate

sorts would be performed, increasing the required tape mounts and potential for
error. If 75,000 cylinders were allocated, most of the input would be processed by
the first intermediate sort, delaying the first breakpoint and introducing the
potential for losing data. It is crucial, therefore, to allocate a balanced amount of
DASD space that will divide your file into reasonably sized segments to minimize
the possibility of system error and to enhance your performance.
Before tuning MAXSORT then, a number of individual environmental elements
should be considered. A study of disk and tape availability, input data file size, and
virtual storage limitations will help you optimize the balance of performance and
reliability.

Syncsort MFX Programmer’s Guide 9–23

9–24 Syncsort MFX Programmer’s Guide
Chapter 10 PARASORT

PARASORT: Parallel Input Processing for Elapsed Time

Improvement
PARASORT improves elapsed time performance for sorts whose input is a multi-
volume tape data set and/or concatenated tape data sets. Reduced elapsed time can
help critical sort applications achieve batch window goals.
The performance improvement from PARASORT is a result of processing the
SORTIN input volumes in a parallel fashion. Depending upon the resources
provided, elapsed time can be reduced up to 20% for 2-way input and up to 33% for
4-way input.
PARASORT requires additional tape units for the application. You will need from
two to eight times the current number of tape units, depending upon resource
availability and the degree of improvement desired. PARASORT automatically
manages the tape units and minimizes the use of the tape drive resources by
deallocating excess tape drives during initialization and releasing all the extra
units at the end of the sort input phase.

PARASORT Applicability
Certain MFX facilities or application characteristics cannot be used with
PARASORT. The following are incompatible with a PARASORT application:
• A SORTIN record format (RECFM) of VS or VBS.
• A SORTIN GDG with a relative reference specified.
• An ASCII tape data set specified for SORTIN.

Syncsort MFX Programmer’s Guide 10–1

 • An exit routine other than a pre-linked or inline E35 exit.
• EQUALS specified either as an installation or run-time option.
• SKIPREC or STOPAFT options specified.
• SEQNUM specified on INREC.
• IFTHEN WHEN=GROUP specified on INREC.
• CKPT (checkpoint) option in effect.
• The MAXSORT option specified.
• Certain unusual sort key types, feature combinations, or long sort keys in
excess of 800 bytes.
• FIELDS=COPY specified.

Job Control Language

The JCL for PARASORT is similar to the JCL for a standard disk sort. The primary
difference is that PARASORT JCL must specify additional tape units to allow
parallel input processing of the SORTIN data set. For details on SORTIN JCL for
PARASORT, see “SORTIN DD Statement with PARASORT” on page 10-4.
To initiate PARASORT using job control statements, specify PARM=PARASORT on
the EXEC statement. A program-initiated sort requests PARASORT by using a
PARM card image in the data set defined by the $ORTPARM DD statement.

Sample EXEC Statement

//stepname EXEC PGM=SYNCSORT, PARM=PARASORT

Figure 387. PARASORT EXEC Statement

10–2 Syncsort MFX Programmer’s Guide

DD Statements
PARASORT’s DD statement requirements are summarized in the following table.
One additional DD type (SORTPARn) is required compared to a conventional disk
sort.

//$ORTPARM DD Used to override PARM or control statement information.

//SYSIN DD Control statement data set. Required unless control statements

are supplied via an invoking program parameter list.

//SYSOUT DD Message data set. Required unless all messages are routed to
console.

//SORTWKxx DD Disk work area definition. Required unless DYNALLOC is spec-

ified.

//SORTIN DD SORT input data set. Required.

//SORTPARn DD Defines additional tape units for parallel reading of SORTIN.

Required.

//SORTOUT DD Output data set. Required unless there is an E35. Ignored if the
invoking program supplies an inline E35 exit routine; optional if
the MODS statement activates an E35 exit routine.

//SORTXDUP DD Output data set for records deleted by DUPKEYS. Required if

XDUP parameter used.

//SORTXSUM DD Output data set for records deleted by SUM. Required if XSUM
parameter used.

//ddname DD Required unless E35 user exit is in LINKLIB/JOBLIB/

STEPLIB.

Table 63. PARASORT DD Statements

The SORTIN and SORTPARn DD statements are discussed below. For a discussion
of the other DD statements, which are specified for PARASORT just as for a non-
PARASORT Disk Sort, see “Chapter 4 JCL and Sample JCL/Control Statement
Streams”.

Syncsort MFX Programmer’s Guide 10–3

SORTIN DD Statement with PARASORT
The SORTIN DD statement is required for a PARASORT application. It must
define either a single multi-volume tape data set or several concatenated tape data
sets, which can be single or multi-volume. If the SORTIN is concatenated, each
data set must meet the normal SORTIN concatenation requirements and must be
able to use the same device type so that UNIT=AFF=SORTIN can be specified for
all data sets in the concatenation. For a discussion of normal SORTIN JCL, see
“Chapter 4 JCL and Sample JCL/Control Statement Streams”
For optimal performance, data sets that reside on tapes, such as 3480s, that can be
read only in a single direction should have two units allocated. If the data set is on
a tape that supports bidirectional processing, a single unit is sufficient. In all cases
DEFER mounting must be specified.
SORTIN data sets may not be passed data sets or have PASS specified on their DD
statement.
Either the catalog or specific list of volume serial numbers must be specified. The
volume serial list must accurately reflect the volumes in the data set. If extra
volumes are specified (as may happen if an old data set is rewritten with less data)
an error message will be generated. A volume sequence number may not be
specified.
The following example SORTIN DD statements for PARASORT illustrate three
different SORTIN cases:
• A multi-volume cataloged data set
• A multi-volume uncataloged data set
• Concatenated single and multi-volume uncataloged data sets
Note that each example includes a y on the UNIT specification (for example,
UNIT=(3480,y,DEFER)). The y is either 1 or 2 and indicates the number of units to
be allocated for these devices. For optimal performance, data sets that reside on
tapes that can be read only in a single direction, such as 3480s, should have two
units allocated. If the data set is on a tape that supports bidirectional processing, a
single unit is sufficient. In all cases DEFER mounting must be specified.
Note also that each example includes an alternative UNIT specification:
UNIT=(xxxxx1,y,DEFER). This specification applies if special esoteric names are
available. The xxxxx1 is a special esoteric unit name established especially for
PARASORT. These esoteric names may have been created at your site for use with
PARASORT. Contact the systems programmer responsible for MFX installation to
determine if they are available. For information on how to select a special esoteric
name, see “Special Channel Separated Esoteric Names” on page 10-7.

10–4 Syncsort MFX Programmer’s Guide

Example 1
SORTIN consists of a single multi-volume cataloged data set.

//SORTIN DD DSN=INPUT.FILE,DISP=(OLD,KEEP),
// UNIT=(,y,DEFER)
or if a special esoteric name is available
// UNIT=(xxxxx1,y,DEFER)

Figure 388. Sample SORTIN DD Statement

Example 2
SORTIN consists of a single multi-volume uncatalogued data set.

//SORTIN DD DSN=INPUT.FILE,DISP=(OLD,KEEP),
// UNIT=(3480,y,DEFER),VOL=SER=(VOL001,VOL002,...,VOL00N)
or if special esoteric name is available
// UNIT=(xxxxx1,y,DEFER),VOL=SER=(VOL001,VOL002,...,VOL00N)

Figure 389. Sample SORTIN DD Statement

This example presumes the file is standard label and the first file is on VOL001.
For uncataloged data sets, the unit and volser list must be specified.

Example 3
SORTIN consists of a concatenation of single and multi-volume uncatalogued data
sets.

//SORTIN DD DSN=INPUT.FILE1,DISP=(OLD,KEEP),
// UNIT=(3480,y,DEFER),VOL=SER=(VOL001,VOL002,VOL003)
or if special esoteric names are available
// UNIT=(xxxxx1,y,DEFER),VOL=SER=(VOL001,VOL002,VOL003)
// DD DSN=INPUT.FILE2,DISP=(OLD,KEEP),
// UNIT=AFF=SORTIN,VOL=SER=(VOL101,VOL102)
// DD DSN=INPUT.FILE3,DISP=(OLD,KEEP),
// UNIT=AFF=SORTIN,VOL=SER=(VOL201)

Figure 390. Sample SORTIN DD Statement

It is also possible to include a DD DUMMY allocation in the SORTIN

concatenation. This is necessary when modifying production JCL that is a
prototype for the maximum number of possible concatenated input files. If a
particular execution uses less than the maximum number, place the DD DUMMY
specification at the appropriate point. This specification should contain a DCB
specification that matches that of the first SORTIN data set.

Syncsort MFX Programmer’s Guide 10–5

SORTPARn DD Statements
The SORTPARn DD statements define units that will be used to perform the
parallel reading of the input file. Up to four SORTPARn DD statements may be
provided, with a minimum of two required. The number of SORTPARn DD
statements that you provide may be limited by the tape channel capacity at your
installation. See “Special Channel Separated Esoteric Names” on page 10-7 for
information on how to determine if your choice is limited.
The n in the SORTPARn is replaced with numbers 1 through 4. The numbers must
start at 1 and be numbered consecutively.
The required SORTPAR1 must be coded in one of the following two ways,
depending on whether the SORTIN data set is cataloged or not.
• If the SORTIN DD is defined as a single cataloged data set or as a series of
concatenated data sets where the first data set of the concatenation is
cataloged, then the SORTPAR1 DD must be coded as follows:

//SORTPAR1 DD DSN=*.SORTIN,DISP=OLD,
// UNIT=AFF=SORTIN

Figure 391. Sample SORTPAR1 DD Statement

• If the SORTIN DD is defined as a single non-cataloged data set or as a series

of concatenated data sets where the first data set of the concatenation is non-
cataloged, then the SORTPAR1 DD must be coded as follows:

//SORTPAR1 DD DSN=*.SORTIN,DISP=OLD,UNIT=AFF=SORTIN,
// VOL=SER=(VOL001,VOL002,...,VOL00n)

Figure 392. Sample SORTPAR1 DD Statement

where the VOL=SER list contains the identical volumes specified on the
SORTIN DD specification. If the SORTIN DD is a series of concatenations, the
VOL=SER list contains the volumes that comprise the first data set in the
concatenation.
The remaining SORTPARnn DDs are coded as shown on the following prototype
SORTPARnn DD statement:

//SORTPARn DD DSN=*.SORTIN,DISP=(,KEEP,KEEP),
// VOL=PRIVATE,UNIT=(xxxx,y,DEFER)
or if special esoteric names are available
// VOL=PRIVATE,UNIT=(xxxxxn,y,DEFER)

Figure 393. Prototype SORTPARn DD Statement

10–6 Syncsort MFX Programmer’s Guide

 The xxxx is a unit type or generic name compatible with the device associated with
SORTIN. If special channel separated esoteric names have been made available,
see “Special Channel Separated Esoteric Names” on page 10-7.
The y is either 1 or 2 and indicates the number of units to be allocated for these
devices. For optimal performance, data sets that reside on tapes that can be read
only in a single direction, such as 3480s, should have two units allocated. If the
data set is on a tape that supports bidirectional processing, a single unit is
sufficient. In all cases DEFER mounting must be specified.
The number of SORTPARn data sets to allocate depends on several factors:
• The total number of volumes to be read from SORTIN and its concatenations.
There is no need to allocate more SORTPARn data sets than total volumes in
the SORTIN file. Note that if more SORTPARn data sets are allocated than
there are volumes, the excess SORTPARn data sets will be deallocated at
PARASORT’s initiation.
• The degree of performance improvement desired.
Typically, two SORTPARn data sets will provide up to 20% elapsed time
improvement; three, up to 25%; and four, up to 33%.
• The degree of channel contention, which may reduce the number of SORTPARn
DD statements used.
The use of special esoteric unit names will ensure that this contention is
eliminated, but your choice for the number of SORTPARn DD statements may
be limited.
• Resource availability.
System constraints may limit the number available to a particular job.

Special Channel Separated Esoteric Names

For optimal PARASORT performance, MFX must be able to read each SORTPARn
input DD simultaneously, with no channel contention. To ensure this, your system
programming staff may have defined special esoteric unit names for use with
PARASORT. Before creating a PARASORT application, you should contact your
system programmer to verify that this work has been done and that the special
names are available for use. Note, however, that even if the work has been done,
certain categories may be unavailable due to limited channel capacity.
To use special esoteric names, do the following:

Syncsort MFX Programmer’s Guide 10–7

 1. Decide whether you would like to specify 4-way (up to SORTPAR4), 3-way (up to
SORTPAR3) or 2-way (up to SORTPAR2) input.

2. In the table of special esoteric names provided by your systems programmer, find the
name that corresponds to the tape type of the SORTIN data sets.

The following is a sample table of special esoteric names. This table is for
illustration only; the names at your site may be different.

3420 3480/90 3490E 3590

|----------|----------|----------|----------|
4-WAY | PAR241 | PAR441 | PARE41 | NOT |
INPUT | PAR242 | PAR442 | PARE42 | POSSIBLE |
| PAR243 | PAR443 | PARE43 | |
| PAR244 | PAR444 | PARE44 | |
|----------|----------|----------|----------|
3-WAY | PAR231 | PAR431 | PARE31 | NOT |
INPUT | PAR232 | PAR432 | PARE32 | POSSIBLE |
| PAR233 | PAR433 | PARE33 | |
|----------|----------|----------|----------|
2-WAY | PAR221 | PAR421 | PARE21 | PAR921 |
INPUT | PAR222 | PAR422 | PARE22 | PAR922 |
|----------|----------|----------|----------|

Figure 394. Sample Esoteric Unit Name Table

3. Use the name from the table at your site for your SORTIN and SORTPARn names. The
following example JCL is for input from 3480 cartridges with at least 4 volumes:

//SORTIN DD DSN=....,DISP=(OLD,KEEP),UNIT=(PAR441,2,DEFER)
//SORTPAR1 DD DSN=*.SORTIN,DISP=OLD,
// UNIT=AFF=SORTIN
//SORTPAR2 DD DSN=*.SORTIN,DISP=(,KEEP,KEEP),
// VOL=PRIVATE,UNIT=(PAR442,2,DEFER)
//SORTPAR3 DD DSN=*.SORTIN,DISP=(,KEEP,KEEP),
// VOL=PRIVATE,UNIT=(PAR443,2,DEFER)
//SORTPAR4 DD DSN=*.SORTIN,DISP=(,KEEP,KEEP),
// VOL=PRIVATE,UNIT=(PAR444,2,DEFER)

Figure 395. Sample JCL

10–8 Syncsort MFX Programmer’s Guide

Sortwork Considerations
The amount of sortwork space required is the same as if the application were run
as a conventional sort. What should be modified, if sortworks are provided via JCL
rather than DYNALLOC, is the number of SORTWKxx DD statements. Try to
provide a total number of SORTWKxx DDs that is two to three times the number of
SORTPARns specified. This would typically require an adjustment in primary and
secondary space amounts so that the total space allocated is similar to that of the
original application. This subdivision of SORTWORK space will provide an
opportunity for additional channel path availability. This parallelism in
SORTWORK channel paths is also a key to improving sort elapsed time
performance.

Operations Notes
Since the SORTIN tape volumes will be read in any order and on a SORTPARn
tape unit location determined by the PARASORT logic, it is possible to receive
messages on the console log that indicate out of sequence processing of the volumes
of the SORTIN data set. Messages such as the following may be generated:
IEC712I ... SORTPARn READ - NOT FIRST VOLUME OF DATA SET
or
IEC710I ... SORTPARn ANOTHER VOLUME EXPECTED
These messages can be disregarded since this type of processing is deliberate with
PARASORT.

Syncsort MFX Programmer’s Guide 10–9

10–10 Syncsort MFX Programmer’s Guide
Chapter 11 MFX DB2 Query Support

MFX can directly retrieve data from a DB2 database based on a user-provided
query. An SQL SELECT statement is used to specify the criteria of the request. The
query of the DB2 database replaces MFX's SORTIN or E15 processing. SORT or
COPY functions, but not MERGE, can be used with DB2 queries. All MFX features
performed after E15 processing are available for use with the DB2 query facility.
Refer to “Chapter 8, The Flow of the Sort” for a summary of MFX's features and
flow of control during processing.
The MFX DB2 Query facility improves performance over DB2’s DSNTIAUL
program by allowing DB2 data to be passed directly into a SORT or COPY
operation, without the use of setup steps or the need for user-written E15 exits.

Restrictions
The following cannot be used with the DB2 Query facility. If specified, they will
cause MFX to terminate with a return code of 16:
• E15 exit
• The SKIPREC parameter
• The MAXSORT feature
• The PARASORT feature
• MERGE
The following will be ignored if used with the DB2 Query facility:
• A SORTIN data set
• The TYPE parameter and the l1 and l2 values of the LENGTH parameter of a
RECORD statement

Syncsort MFX Programmer’s Guide 11–1

Job Control Language
The JCL for the DB2 Query facility is similar to the JCL of a standard disk sort.
The primary difference is that the DB2 query JCL must also contain an additional
SORTDBIN DD specification to define the DB2 query with an SQL SELECT
statement.
To initiate a SORT or COPY with the DB2 Query facility using job control
statements, specify PARM='DB2=dsn' on the EXEC statement. The dsn referred to
in the DB2 parameter is the DB2 subsystem name to be accessed. When a SORT or
COPY DB2 Query application is invoked from a program, specify the DB2
parameter in the $ORTPARM DD statement.
Note: In order to issue the first query to the DB2 subsystem identified in the
DB2=parm, you must have BINDADD authority so the SYNCSORT packages and
plan can be added to the subsystem. If you don’t want to use the default DB2
options to bind our SYNCSORT packages and plan, you can manually do the bind-
ing by using DBRM members which are included with your MFX installation files.
Contact Precisely Support for additional instructions.

Sample EXEC Statement

The following shows a sample EXEC statement with the DB2 PARM:

//stepname EXEC PGM=SYNCSORT, PARM='DB2=dsn'

Figure 396. DB2 Query EXEC Statement

DD Statements
The DD statements used with the DB2 Query facility are summarized in this table.
Note that the SORTDBIN DD statement is unique to the DB2 Query facility.

//$ORTPARM DD Used to override PARM or control statement information.

//SYSIN DD Control statement data set. Required unless control statements

are supplied via an invoking program parameter list.

//SYSOUT DD Message data set. Required unless all messages are routed to
console.

//SORTWKxx DD Disk work area definition. Required unless DYNALLOC is spec-

ified.

Table 64. (Page 1 of 2)DB2 Query DD Statements

11–2 Syncsort MFX Programmer’s Guide

 //SORTDBIN DD Data set which defines the DB2 query. Contains the SQL
SELECT statement to be used for this application.

//SORTOUT DD Output data set. Required unless there is an E35. Ignored if the
invoking program supplies an inline E35 exit routine; optional if
the MODS statement activates an E35 exit routine.

//SORTXDUP DD Output data set for records deleted by DUPKEYS. Required if

XDUP parameter used.

//SORTXSUM DD Output data set for records deleted by SUM. Required if XSUM
parameter used.

//SORTMODS DD Required if user exits are in SYSIN and if user exits are to be
//SYSLIN DD linkage-edited at execution time.
//SYSLMOD DD
//SYSPRINT DD

//ddname DD Required for exits unless the exit is inline in LINKLIB/JOBLIB/

STEPLIB or in SYSIN.

Table 64. (Page 2 of 2)DB2 Query DD Statements

SORTDBIN DD Statement
The SORTDBIN DD statement is required for a DB2 query application. The data
set defined by the SORTDBIN contains the SQL SELECT statement that describes
the criteria of the query.
The SORTDBIN DD record format must be F or FB, and the record length must be
80.
The SORTDBIN data set must be formatted in accordance with the following rules:
• Only a SELECT statement or $ELECT statement (for trial run described below)
is accepted. Any other SQL statements will cause the job to terminate with a
WER468A error message. For details on the facilities and syntax of a SELECT
statement refer to the IBM publication DB2 Universal Database for z/OS SQL
Reference (SC18-7426).
• The maximum supported length of a SELECT statement for this feature is
32765 characters.
• The SELECT statement may not use the '--' convention of two consecutive
hyphens to denote that the remainder of a card image is a comment.
• The SELECT statement may be terminated with a semicolon. Any characters
found after the semicolon will be considered comments.
• Only columns 1 through 72 of each record will be read. Columns 73 through 80
will be ignored.

Syncsort MFX Programmer’s Guide 11–3

PARM Options
The DB2 query support parameters described below may be specified on the EXEC
statement, in the $ORTPARM data set, in PARMTBLE or in a PARMEXIT. They
may be listed in any order.

DB2

DB2=dsn

Figure 397. DB2 Format

This parameter is required in order to execute MFX’s DB2 Query facility. The dsn
referred to in the DB2 parameter is the DB2 subsystem name to be accessed.

MULTIFETCH

MULTIFETCH = { n, 100 }

Figure 398. MULTIFETCH Format

This parameter indicates the number of rows per multiple-row fetch. n can be any
integer from 1 to 1000. The default 100 rows per fetch will be used if
MULTIFETCH is not specified. A higher number of rows per fetch may enhance
DB2 query performance, but more storage will be consumed. When the storage
amount exceeds an internal value, MFX will dynamically reduce the number of
rows per fetch.
The multiple-row fetch feature can only be used for DB2 version 8 and above. It will
be ignored for prior DB2 versions.

Operation
Using the query provided in the SORTDBIN data set, MFX will access the DB2
database specified in the DB2 EXEC PARM and will process as fixed-length records
the rows returned from the query. The records will be processed as if read from a
SORTIN or retrieved from an E15 exit. All MFX features available after E15
processing in the flow of control can be used with a SORT or COPY application.
The record used within MFX is constructed from the fields in the query as follows:
• The order of the fields in the record is the same as the order specified by the
SELECT statement.

11–4 Syncsort MFX Programmer’s Guide

 • The data format of the fields within the record is the same format returned by
DB2.
• A fixed-length field is the same length as returned by DB2.
• Variable-length character data is stored in a fixed-length field. The field length
is equal to the maximum length of the field plus two bytes for a leading field
length descriptor variable. The field length descriptor contains a binary value
describing the number of bytes of data provided for this field. If an instance of
the field is shorter than the maximum, the remaining bytes will be set to binary
zeros.
• Any fields defined to allow nulls will cause the creation of two fields within the
record constructed by MFX. The first will be the data field and the second will
be a one byte indicator field. If the value of the field is null, by default the field
will be filled with binary zeros (X'00') and the indicator field will contain a '?' to
signify the field is null. If the value of the field is not null, the indicator will be
set to binary zeros. If binary zeros would not be an appropriate fill value for a
null field, use of SQL functions such as VALUE or COALESCE on the SELECT
statement should be considered. For instance, if the field to be retrieved is
packed decimal, it is usually best to create a null value of the proper PD format.
This ensures that if the field is used later as a sort key or in other data
conversion features, it would contain appropriate high or low values such as PD
zeros or nines as specified in the VALUE or COALESCE function.

Record Description
Information about the record created by the query will be displayed in the MFX
message data set. For each column selected, the report will display the start and
end position within the record, the DB2 data type, the equivalent MFX data type,
and whether null values are allowed. A data type whose length is not implied by the
format will have a field length appended to its description. Note that the length
displayed for a VARCHAR DB2 data type is two bytes shorter than would be
indicated by the field start and end positions. The extra two bytes described in the
start and end positions are for the field length descriptor, which is contained in the
first two bytes of the field.

Syncsort MFX Programmer’s Guide 11–5

Record Description: Trial Mode Execution
When first developing an application, knowledge of the actual input record layout
built from the query is required. This can be obtained from a trial mode execution.
The trial mode execution uses a query provided in the SORTDBIN data set to
generate a report of the input record layout in the MFX message data set
(SYSOUT). The trial mode execution does not perform any other processing or
request data from DB2. To request trial mode execution, modify the SQL SELECT
keyword in the SORTDBIN data set to $ELECT. This indicates a trial mode
execution is to be performed. For trial mode, execute MFX with JCL of the following
form:

// EXEC PGM=SYNCSORT,PARM='DB2=DSN1'
//STEPLIB DD DSN=DB2.SDSNLOAD,DISP=SHR
// DD DSN=SORT.RESI.DENCE,DISP=SHR
//SYSOUT DD SYSOUT=A
//SORTDBIN DD *
$ELECT FIRSTNME,LASTNAME,WORKDEPT,HIREDATE,EDLEVEL,SALARY
FROM DSN.EMPTAB
WHERE EDLEVEL>10
/*

Figure 399. Sample JCL for Trial Mode Execution

Note that only the SYSOUT and SORTDBIN DD are required for a trial mode
execution. An actual execution of the application will require other DDs as
documented for SORT or COPY applications.
The STEPLIB DD statement specifies where the MFX and DB2 products can be
found. The STEPLIB DD statement would be needed if these products could not be
found in the standard system libraries.
The DB2 EXEC statement parameter would be set to the DB2 subsystem name to
be accessed.
The SYSOUT data set will contain an input record layout report as shown in the
following sample:

11–6 Syncsort MFX Programmer’s Guide

 .
.
.
DB2 QUERY OPTION SELECTED
QUERY STATEMENTS:
SELECT FIRSTNME,LASTNAME,WORKDEPT,HIREDATE,EDLEVEL,SALARY
FROM DSN.EMPTAB
WHERE EDLEVEL>10
INPUT RECORD DESCRIPTION:

COLUMN START END DB2 SYNCSORT NULL VALUE

NAME POSITION POSITION DATA TYPE DATA TYPE
---------- -------- -------- -------------------- --------- ----------
FIRSTNME 1 14 VARCHAR(12) CH DISALLOWED
LASTNAME 15 31 VARCHAR(15) CH DISALLOWED
WORKDEPT 32 34 CHAR(3) CH ALLOWED
NULLINDICATOR 35 35 CH
HIREDATE 36 45 DATE CH ALLOWED
NULLINDICATOR 46 46 CH
EDLEVEL 47 48 SMALLINT BI ALLOWED
NULLINDICATOR 49 49 CH
SALARY 50 54 DECIMAL(9,2) PD ALLOWED
NULLINDICATOR 55 55 CH

WER467I DB2 QUERY TRIAL MODE SUCCESSFULLY EXECUTED

.
.
.

Figure 400. Sample SYSOUT

You would create MFX control statements with field specifications based on the
input record layout and place the control statements in the data set specified by the
SYSIN DD statement. You would then create a set of JCL statements for the
application.

Syncsort MFX Programmer’s Guide 11–7

Sample MFX DB2 Query Application
In this example, a query is made for employee data. The example specifies how the
application is to sort and format the data into a report. The formatting adds
headers, field spacing, and converts a date to printable forms.

//SORTSQL EXEC PGM=SYNCSORT,PARM='DB2=DSN1'

//STEPLIB DD DSN=DB2.SDSNLOAD,DISP=SHR
// DD DSN=SORT.RESI.DENCE,DISP=SHR
//SYSOUT DD SYSOUT=A
//SORTOF1 DD DSN=OUT1,DISP=(NEW,CATLG),UNIT=3390,SPACE=(CYL,1)
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,20)
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,20)
//SORTDBIN DD *
SELECT FIRSTNME,LASTNAME,WORKDEPT,HIREDATE,EDLEVEL,SALARY
FROM DSN.EMPTAB
WHERE EDLEVEL>10
//SYSIN DD *
SORT FIELDS=(3,12,CH,A)
OUTFIL FILES=1,
HEADER1=(2/,20:'EMPLOYEE INFORMATION',
2/,1:'FIRSTNAME',14:'LASTNAME',29:'WORKDEPT',
41:'HIRE DATE',54:'LEVEL',65:'SALARY',
/,1:'---------',14:'--------',29:'---------',
40:'----------',54:'------',65:'---------'),
OUTREC=(1:3,12,C' ',17,15,C' ',32,3,7C' ',
36,10,C' ',47,2,BI,M0,5C' ',50,5,PD,M2)

Figure 401. Sample MFX DB2 Query Application

The following describes the JCL statements:

• The EXEC statement identifies SYNCSORT as the program to be executed. The
DB2 PARM defines the DB2 subsystem to be accessed.
• The STEPLIB DD statement instructs the system as to where the MFX and
DB2 products can be found.
• The SYSOUT DD statement assigns the MFX messages to the output device
associated with SYSOUT class A.
• The SORTOF1 DD statement gives OUT1 as the output data set name and
specifies a 3390 disk. One cylinder of primary space has been allocated on this
volume. The DISP parameter shows that this data set is not yet in existence.
• The two SORTWK statements reserve space on four temporary data sets for
inter-mediate storage. Twenty cylinders are to be reserved on the data sets.
• The SORTDBIN DD statement marks the beginning of the input stream that
contains the SQL SELECT statement that describes the criteria of the query.

11–8 Syncsort MFX Programmer’s Guide

 • The SYSIN DD statement marks the beginning of the input stream that
includes the sort control statements. A sort will be performed and a report will
be generated. The records read from the DB2 database under control of the
query specified in the SORTDBIN data set will be formatted and presented in
this report. Fields will be converted to printable format when necessary.
The SYSOUT will contain a report on the execution of the application. The report
displays the control statements followed by the query record layout and MFX
messages with information on the particular execution. The following is a sample
report:

.
.
.
SYSIN :
SORT FIELDS=(3,12,CH,A) 00048000
OUTFIL FILES=1, 00049002
HEADER1=(2/,20:'EMPLOYEE INFOMATION', 00050002
2/,1:'FIRSTNAME',14:'LASTNAME',29:'WORK DEPT', 00051002
41:'HIRE DATE',54:'LEVEL',65:'SALARY', 00052002
/,1:'---------',14:'--------',29:'---------', 00053002
40:'----------',54:'------',65:'---------'), 00054002
OUTREC=(1:3,12,C' ',17,15,C' ',32,3,7C' ', 00060002
36,10,C' ',47,2,BI,M0,5C' ',50,5,PD,M2) 00070002
DB2 QUERY OPTION SELECTED
QUERY STATEMENTS:
SELECT FIRSTNME,LASTNAME,WORKDEPT,HIREDATE,EDLEVEL,SALARY
FROM DSN.EMPTAB
WHERE EDLEVEL>10
INPUT RECORD DESCRIPTION:

COLUMN START END DB2 SYNCSORT NULL VALUE

NAME POSITION POSITION DATA TYPE DATA TYPE
---------- -------- -------- -------------------- --------- ----------
FIRSTNME 1 14 VARCHAR(12) CH DISALLOWED
LASTNAME 15 31 VARCHAR(15) CH DISALLOWED
WORKDEPT 32 34 CHAR(3) CH ALLOWED
NULLINDICATOR 35 35 CH
HIREDATE 36 45 DATE CH ALLOWED
NULLINDICATOR 46 46 CH
EDLEVEL 47 48 SMALLINT BI ALLOWED
NULLINDICATOR 49 49 CH
SALARY 50 54 DECIMAL(9,2) PD ALLOWED
NULLINDICATOR 55 55 CH
.
.
.

Figure 402. Sample SYSOUT Report

Syncsort MFX Programmer’s Guide 11–9

 The following shows the output from the application:

EMPLOYEE INFOMATION

FIRSTNAME LASTNAME WORK DEPT HIRE DATE LEVEL SALARY

--------- -------- --------- ---------- ------ ---------
CHRISTINE HAAS A00 01/01/1975 18 52,750.00
CHRISTINE MIKE A00 06/08/1978 12 53,330.00
DIANE HARISON A00 02/01/1978 13 12,500.00
DIANE HEMMINGER A00 01/01/1975 14 46,500.00
JOAN PAN A00 05/01/1973 15 12,110.00
KEVEN MASK B00 06/01/1988 14 34,780.00
MAGGIE NEME A00 06/01/1978 13 54,330.00
MIKE BUSH B00 12/08/1987 11 12,340.00
PETER MAWAH B00 02/01/1988 18 30,000.00
STEVE ARNEY B00 02/01/1990 18 34,560.00

Figure 403. Sample Application Output

11–10 Syncsort MFX Programmer’s Guide

Chapter 12 Multiple Input Files

MFX can process multiple VSAM and non-VSAM data sets for input through the
MULTIIN facility. This facility enhances the standard SORTIN specification
which supports only a single VSAM data set for input without any concatenation
of VSAM or non-VSAM data sets. The MULTIIN facility can be used for a SORT or
COPY application.
MULTIIN may also be used in any multiple input file application where there is a
need to identify the data set from which a record was read. The &MULTIINDD
subparameter of the INCLUDE/OMIT control statements and the INREC control
statement can be used for this purpose. See “INCLUDE/OMIT Control Statement
Format” on page 2-29 and “INREC Control Statement Format” on page 2-55 for
details.
If multiple files need to be processed as input, but there are no VSAM files and
there is no need for the &MULTIINDD feature, conventional SORTIN
concatenation should be used. This will improve the performance of the application.

Syncsort MFX Programmer’s Guide 12–1

Restrictions
The MULTIIN facility cannot be used in a MERGE or JOIN application. If
specified, the MFX application will terminate with a return code of 16.

Job Control Language

The JCL for the MULTIIN facility is similar to the JCL for a standard sort or copy.
The primary difference is that the MULTIIN JCL must contain SORTMInn DD
specifications in place of the SORTIN DD specification. To initiate a SORT or
COPY with the MULTIIN facility using job control statements, specify
PARM=’MULTIIN’ on the EXEC statement. When a SORT or a COPY MULTIIN
application is invoked from a program, specify the MULTIIN parameter in the
$ORTPARM data set.

EXEC Statement
The following shows a sample EXEC statement:

//stepname EXEC PGM=SYNCSORT , PARM=MULTIIN

Figure 404. MULTIIN EXEC Statement

DD Statements
The DD statements used with the MULTIIN facility are summarized in the
following table. Note that the MULTIIN DD statement is unique to the MULTIIN
facility.
//$ORTPARM DD Used to override PARM or control statement information.
//SYSIN DD Control statement data set. Required unless control statements
are supplied via an invoking program parameter list.
//SYSOUT DD Message data set. Required unless all messages are routed to
console.
//SORTWKxx DD Disk work area definition. Required unless DYNALLOC is spec-
ified.
//SORTMInn DD Data sets that define the input for a SORT or COPY application.
Required for a MULTIIN application. Ignored if an invoking pro-
gram supplies an inline E15 exit routine.
//SORTOUT DD Output data set. Required unless there is an E35. Ignored if the
invoking program supplies an inline E35 exit routine; optional if
the MODS statement activates an E35 exit routine.

12–2 Syncsort MFX Programmer’s Guide

 //SORTXDUP DD Output data set for records deleted by DUPKEYS. Required if
XDUP parameter used.
//SORTXSUM DD Output data set for records deleted by SUM. Required if XSUM
parameter used.
//SORTMODS DD Required if user exits are in SYSIN and if user exits are to be
//SYSLIN DD linkage-edited at execution time.
//SYSLMOD DD
//SYSPRINT DD
//ddname DD Required for exits unless the exit is inline in LINKLIB/JOBLIB/
STEPLIB or in SYSIN.

Table 65. MULTIIN DD Statements

SORTMInn DD Statements
SORTMIn or SORTMInn DD statements are required to define the input to a
multiple input application. A SORTIN DD statement will be ignored.
SORTMIn and SORTMInn data sets can be VSAM (entry-sequenced, key-
sequenced or relative record) or non-VSAM data sets, including BatchPipes, z/OS
pipes and HFS data sets. DCB information need not be supplied for a disk or
standard labeled tape file. Any of the information accessed from a standard label
can be overridden by coding the appropriate DCB parameter in the JCL.
It is possible to sort or copy up to 100 data sets. Each input data set is specified
on a SORTMIn or SORTMInn DD statement. The valid range for n is 0 through
9; for nn, 00 through 99. If both SORTMIx and a SORTMI0x are specified, they
are treated as duplicates and only the first definition is processed. Numbers may
be skipped or used out of order. MFX will read the SORTMI files in numerical
order. If EQUALS is in effect, the order of equal-keyed records within each
SORTMInn file will be preserved. In addition, equal-keyed records from the
lowest-numbered SORTMInn file will be written before those from the second
SORTMInn file, and so on.
If no SORTMInn or SORTMIn data sets are defined when the “MULTIIN” PARM is
passed, error message WER224A will be posted.
SORTMInn or SORTMIn cannot have concatenated input files. If a concatenation
is present, error message WER509A will be issued.
The maximum record lengths supported are 32,760 bytes for fixed-length records
and 32,767 bytes for variable-length records.
By default MFX does not accept an uninitialized SORTMInn data set and will
terminate processing with a WER400A message. An uninitialized data set is one
that has been newly created but never successfully closed. The UNINTDS PARM or
installation option can be used to change MFX’s default mode of processing to
accept an uninitialized input data set and process it as an empty file. See
“UNINTDS” on page 5-30.

Syncsort MFX Programmer’s Guide 12–3

 The following shows sample JCL for multiple input data sets:

//SORTMI17 DD DSN=AUGUST.SALES.KSDSDA,DISP=(OLD,KEEP),UNIT=3390
//SORTMI1 DD DSN=JUNE.SALES,DISP=(OLD,KEEP),UNIT=3390,
// VOL=SER=242424,DCB=(LRECL=200,RECFM=VB,BLKSIZE=8004)
//SORTMI14 DD DSN=JULY SALES,DISP=(OLD,KEEP),
// UNIT=TAPE,VOL=SER=654321,LABEL=(1,SL),
// DCB=(LRECL=100,RECFM=VB,BLKSIZE=8004)

Figure 405. Sample SORTMInn DD Statements (multiple input)

In the preceding example, SORTMI17 is a VSAM data set, SORTMI1 is a non-

VSAM disk data set and SORTMI14 is a non-VSAM tape data set.

Operation
Record TYPE Determination
If there are one or more non-VSAM SORTMInn input data sets, the non-VSAM
input RECFM will be used as the input RECFM for the application. All non-
VSAM input files must have the same record format. If the RECORD TYPE
specification differs from the SORTMInn RECFM DCB parameter for the non-
VSAM input, the latter takes precedence.
If all SORTMInn files are VSAM, then the TYPE parameter on the RECORD
control statement should be specified. If TYPE is not provided, the SORTOUT
RECFM will be examined to determine the input TYPE. If no SORTOUT RECFM
is found and the input files are all VSAM, TYPE=V will be assumed if the
SORTOUT is VSAM and TYPE=F will be assumed if SORTOUT is non-VSAM or
there is no SORTOUT.

Record Length Determination

If the record TYPE determined above is variable, then the record length used will
be the largest of the provided files.
If the record TYPE is fixed and one or more SORTMInns are non-VSAM, then the
record length of the non-VSAM data sets will be used. Any VSAM file with a record
shorter than the non-VSAM record length will have those records padded with a
X’00’. If there is a VSAM data set with a record whose length is longer than the
non-VSAM record length, error message WER264A will be issued.
If the record TYPE is fixed and all SORTMInn data sets are VSAM, then the record
length will be the largest record length of all the VSAM data sets. Each record in a
VSAM file whose record length is not the largest one will have its records padded
with X’00’s.

12–4 Syncsort MFX Programmer’s Guide

Notes
If VSAMEMT=YES is specified, any empty VSAM input data set will be processed
as a legitimate data set containing 0 records, and MFX will end with a return code
of 0. The delivered default, VSAMEMT=NO, instructs MFX to terminate with a
WER254A critical error if there is an empty VSAM input data set specified for
input.

Sample Multiple Input File JCL and Control Statements

The following example illustrates how the JCL could be coded for a typical multiple
input application.

//MINAPP JOB Gives the Jobname

// EXEC PGM=SYNCSORT, Identifies the Program
// PARM=’MULTIIN’ Requests MULTIIN
//SYSOUT DD SYSOUT=* Assigns MFX Messages
//* to I/O Device
//*
//SORTMI17 DD DSN=AUGUST.SALES.KSDSDA, Defines One Input
// DISP=(OLD,KEEP),UNIT=3390
//SORTMI1 DD DSN=JUNE.SALES,DISP=(OLD,KEEP), Defines One Input
// UNIT=3390,VOL=SER=242424,
// LRECL=200,RECFM=VB,BLKSIZE=8004
//SORTMI14 DD DSN=JULY.SALES,DISP=(OLD,KEEP),
// LRECL=100,RECFM=VB,BLKSIZE=8004, Defines One Input
// UNIT=TAPE,VOL=SER=654321,
// LABEL=(1,SL)
//SORTOUT DD SYSOUT=* Defines Output Data Set
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,15) Defines Intermediate
//* Storage
//SYSIN DD *
RECORD TYPE V Record Type Statement
SORT FIELDS=(1,14,CH,A) Sorts Records

Figure 406. JCL and Required Control Statements for a Multiple Input Application

Syncsort MFX Programmer’s Guide 12–5

12–6 Syncsort MFX Programmer’s Guide
Chapter 13 The Dictionary Feature

The MFX Dictionary Feature allows you to create symbolic ‘dictionary_names’ for
fields, constants, or output columns, and use these dictionary_names in MFX
control statements.
Using the Dictionary Feature has many benefits:
• Easier coding of control statements speeds development of applications and
improves accuracy.
• More readable and understandable control statements facilitate debugging and
adaptation to changes in the future.
• Reducing or eliminating changes to control statements when record layouts
change saves time and reduces errors.
To use the MFX Dictionary Feature, do the following:
• Build a symbols dictionary using the dictionary statements. A dictionary
statement creates a dictionary_name and associates it with a constant value,
field specification (position, length, format), or a parsed field.
• Activate the Dictionary Feature using the SYMNAMES DD statement. This DD
tells MFX the name and location of the symbols dictionary to use. You can
include multiple dictionaries by concatenating data sets.
• Use dictionary_names in control statements.
You may also use JCL statements with SET symbols or PROC symbols to establish
dictionary_names. See “Using JCL SET and PROC Symbols to Create
Dictionary_Names” on page 13-29.
A dictionary consists of one or more dictionary statements. A dictionary can be a
sequential data set, a member of a PDS, or it can be placed immediately after the
SYMNAMES DD statement (DD *).
Each time MFX executes with the Dictionary Feature activated by the
SYMNAMES DD, it reads the dictionary statements from the symbols dictionary,

Syncsort MFX Programmer’s Guide 13–1

 and substitutes any dictionary_name found in the following control statements
with the associated field specification or constant.
When a record layout changes, just modify the dictionary statements. The next sort
execution will read in the updated dictionary statements and apply the new values
during dictionary_name substitutions.

Building a Dictionary
There are three types of dictionary statements:
• The constant_name statement
• The field_name statement
• The operator statement

Constant_names and field_names associate symbols with constants or field

definitions; operators (POSITION, SKIP and ALIGN) are used to position fields you
are defining.
In addition, you can have a comment statement or a blank statement:
• A statement with an asterisk (*) in the first column is a comment statement.
MFX will not process it, but it can be printed.
• A blank statement contains blanks in positions 1 through 80. MFX will not
process it, but it can be printed.
The following general rules apply to dictionary statements:
• A dictionary statement can start in any column from 1 through 80. However, a
dictionary statement can only occupy one line; a continuation to a second line is
not permissible.
• One or more blanks after the value indicate the beginning of a comment.
Characters following the blank(s) are not processed by MFX but can be printed.
• A semicolon (;) can be used instead of a comma (,) to separate the
dictionary_name and the value.

13–2 Syncsort MFX Programmer’s Guide

Dictionary Statement Format
The format of the dictionary statement is illustrated below.

 constant_name,constant 
 
  ,p,l,f  
   
  ,p,l  
 field_name    [ comment ]
  ,p  
  ,%pp  
   
 
 operator,value 

Figure 407. Dictionary Statement Format

The following rules apply to constant_names and field_names:

• May consist of 1 through 50 EBCDIC characters.
• May be a combination of uppercase letters (A-Z), lowercase letters (a-z),
numbers (0-9), the number sign (#), the dollar sign ($), the commercial at sign
(@), the underscore (_), or the hyphen (-).
• May not have a number (0-9) or a hyphen (-) as the first character.
• Are case-sensitive; for example, 'Address', 'ADDRESS', and 'address' are treated
as different fields.
• Cannot be an MFX reserved word. The following table lists the MFX reserved
words:

Syncsort MFX Programmer’s Guide 13–3

 •

A DATEDIFF M0 through M9 TE1

AC DATENS M00 through M99 TE2
ADD DC1 NEXTDxxx TE3
ADDDAYS DC2 (xxx={SUN,MON,...}) TE4
ADDMONS DC3 NONE TIME
ADDYEARS DC4 NUM TIME1
ALIGN DE1 OL TIME1P
ALL DE2 ONLY TIME2
AND DE3 OR TIME2P
AQ DE4 OT TIME3
ASL DIV PAGE TIME3P
AST DT PAGEHEAD TIMENS
AVG DT1 PD TM1
BI DT2 PD0 TM2
CH DT3 PDC TM3
CLO DTNS PDF TM4
COPY DTNSP POSITION TS
COUNT E PREVDxxx UFF
COUNT15 F (xxx={SUN,MON,...}) UNPAIRED
CSF F1 PSI VALCNT
CSL F2 PZ VLEN
CST FD SEQNUM X
CTO FI SFF XDUP
D FL SKIP XSUM
D1 FS SORTED Y2x*
D2 H SS Y2xx*
DATE HEX SUB Y4x*
DATE1 LASTDAYx SUBCOUNT Z
DATE1P (x={M,Q,W,Y}) SUBCOUNT15 ZD
DATE2 LS SUBDAYS ZDC
DATE2P MAX SUBMONS ZDF
DATE3 MIN SUBYEARS ZSI
DATE3P MOD TC1
DATE4 MUL TC2
DATE5 MULTIINDD TC3
TC4

* x represents any character

Table 66. MFX Reserved Words

All MFX reserved words are in uppercase. Thus, mixed case and lowercase forms of
MFX reserved words are permissible as dictionary_names. Similarly, dictionary
statement operators POSITION, SKIP and ALIGN are in uppercase; thus mixed
and lowercase forms of these words are permissible as dictionary_names. For
example, ‘position’, ‘Skip’, and ‘align’ are all acceptable as dictionary_names.

13–4 Syncsort MFX Programmer’s Guide

 The following sections describe the three types of dictionary statements in detail:
• “The Constant_name Statement: Rules and Syntax” on page 13-5.
• “The Field_name Statement: Rules and Syntax” on page 13-9.
• “The Operator Statement: Rules and Syntax” on page 13-15.

The Constant_name Statement: Rules and Syntax

The format of the constant_name statement is illustrated below.

constant_name,constant [ comment ]

Figure 408. Constant_name Statement Format

The value of a constant may be a character string, a decimal number, a

hexadecimal string, a bit string, a two-digit year date string, or a system symbol
string.
A constant_name representing a specific value can be used whenever it is valid on
an MFX control statement. Note that the length and format must be compatible
with the usage on the control statement.
The table starting on the next page describes the types of constant values.

Syncsort MFX Programmer’s Guide 13–5

 Value Type Description Valid Examples Invalid Examples

Valid formats: '+0.245' C'CITY''

'xx...x' c'BOOK' (unnecessary extra apostro-
C'xx...x' C'O''DOOLE' phe after Y)
nC'xx...x'
c'xx...x' c'city
nc'xx...x' (missing ending apostrophe)

where x is an EBCDIC
character and n is a repeti-
tion factor for the string.
Character string
The maximum length for n
is 4095. The maximum
length of the string is 64
characters.

To include a single apostro-

phe (') in a character string,
use two single apostrophes
(''), which count as 2 charac-
ters.

Valid formats: +100 ++100

n 100 (too many plus signs)
+n -500
-n 00049 500-
Decimal number (minus sign in wrong place)
Maximum length is 31 sig-
nificant digits. Decimal 5.5
points are invalid. (decimal point not allowed)

Valid formats: X'F1C5' X'F1H5'

X'yy...yy' x'3fb91e' (H is not a valid hexadecimal
nX’yy...yy’ X'06' digit)
x'yy...yy'
nx’yy...yy’ x'cf2'
(unpaired hexadecimal digit 2)
where yy represents any
Hexadecimal string pair of hexadecimal digits
and n is a repetition factor.
The maximum value for n is
4095. The maximum length
of the string is 32 pairs of
hexadecimal digits. Hexa-
decimal digit are 0-9, A-F,
or a-f.

Table 67. (Page 1 of 2) Types of Constant Values

13–6 Syncsort MFX Programmer’s Guide

 Value Type Description Valid Examples Invalid Examples

Valid formats: b'11110000' b'0011'

B'bbbbbbbb...bbbbbbbb' B'11...0000' (only 4 bits, 8 needed)
b'bbbbbbbb...bbbbbbbb'. b'01101111'
b''
Each group of 8 bs rep- (no bits specified)
Bit string resents the 8 bits that com-
pose one byte. Maximum B'00001112'
length is 8 groups of 8 bits (invalid bit value 2)
each. A bit is a 1, 0 or .
(period). Must be a multiple
of 8 bits.

Valid formats: Y'HIGH' Y'1'

Y'LOW' Y'LOW' (fewer than 2 digits)
Y'HIGH' Y'980731'
Y'BLANKS' Y'012' Y' '
Y'x...x' (blank is not a valid digit)
Y'DATE1'
Y'1234567'
Y'DATE2'
(more than 6 digits)
Y'DATE3'
Two-digit year date y'LOW' Y'BLANK'
string y'HIGH' (should be 'BLANKS')
y'BLANKS'
y'x...x'
y'DATE1'
y'DATE2'
y'DATE3'

where x...x represents 2 to 6

decimal digits.

Valid formats: S'&YYMMDD' S'&hhmmss'

S'&xx...xx' s'&SYSNAME' (not in uppercase)
s'&xx...xx' s'&JOBNAME'
System Symbol S'&SYSPLEX'
where x...x represents the
system symbol all in upper-
case.

Table 67. (Page 2 of 2) Types of Constant Values

Specifying System Symbols

System symbols are defined in IBM publication SA22-7592 z/OS MVS
Initialization and Tuning Reference. You can use dynamic system symbols, system-
defined static system symbols and installation-defined static system symbols in
your symbol dictionary and MFX control statements.
The string you specify as a symbol string can consist of a mixture of characters and
all three types of system symbols. You can also build strings using concatenation
and substrings. If you need to specify an apostrophe in the string, use two single
apostrophes.

Syncsort MFX Programmer’s Guide 13–7

 System symbols must be specified in uppercase.
When system symbols are encountered in a symbol dictionary, they are converted
to the appropriate character string and then processed as any other symbol.

13–8 Syncsort MFX Programmer’s Guide

The Field_name Statement: Rules and Syntax
The format of the field_name statement is illustrated below.

 ,p,l,f 
 
 ,p,l 
field_name   [ comment ]
 ,p 
 ,%pp 
 

Figure 409. Field_name Statement Format

A field_name in a dictionary statement can be defined by its position in the record

and its length and format type (p,l,f). Length and format type are optional. A
field_name can also refer to a parsed field (%pp).
The rules for fields specified in a control statement always apply to a field specified
as a dictionary_name since internally MFX substitutes the actual field
specification. For example, if you specify a format, take care that the format of the
field_name is acceptable on the MFX control statement. For example, consider the
dictionary_name CITY, defined by the following dictionary statement:

CITY,10,29,CH

Figure 410. Example of field_name in Dictionary Statement

The following control statement specifying CITY would be valid because a CH field
is permissible on the SORT control statement:

SORT FIELDS=(CITY,A)

Figure 411. Example of field_name in Control Statement

However, the field_name CITY could not be used on a DUPKEYS AVG control
statement because CH is not a valid format with AVG.
Note that you can specify a field_name with p, l, and f, then use the field_name in a
control statement that only requires p and l. MFX will substitute p and l during
processing and ignore the f specification. For example, consider the following
dictionary statements:

Account#,1,12,CH
CheckAmount,23,6,PD

Figure 412. Example of p,l,f in Dictionary Statements

Syncsort MFX Programmer’s Guide 13–9

 Suppose you use these field_names in the following control statements:

SORT FIELDS=(Account#,A)
DUPKEYS SUM=(CheckAmount),FORMAT=PD

Figure 413. Example of field_names in Control Statements

Based on the field_names defined in the dictionary statements, MFX will make the
following substitutions:
• In the SORT statement, 'Account#' is replaced with '1,12,CH'.
• In the DUPKEYS statement, 'CheckAmount' is replaced with '23,6'.
Here are the resulting control statements:

SORT FIELDS=(1,12,CH,A)
DUPKEYS SUM=(23,6),FORMAT=PD

Figure 414. Example of p,l in Control Statements

For more information on format substitution, refer to the “Using Dictionary_names

in MFX Control Statements” section starting on 13-20.
The following three subsections describe the rules for specifying position (p), length
(l) and format (f) in field_name dictionary statements.

Specifying Position (p) in Field_name Dictionary Statements

Following are the rules for specifying position (p) in field_name dictionary
statements:
• p can be a number from 1 through 32752 whenever p,l or p,l and f are used.
However, if position (p) is used alone (for example, CITY,20), p can not be more
than 31 significant digits. Note that any value of p greater than 32752 may
cause a syntax error when it is processed as a position.
• p can be in the form of m.n for specifying a bit position, where m is a number
from 1 through 32752 and n is a number from 0 through 7. Note that m.0 is
equivalent to m.
• p can be an asterisk (*), which indicates that the next position should be
assigned to p. Since l represents length, the next position is set to p + l each
time the field_name for p,l,f or p,l is encountered. If the next position has not
yet been determined, as when the asterisk is used in the first field_name, then
p defaults to one.
When p is an asterisk (*) and there is no length specified, you can also specify +
or – n, where n is a number from 1 through 32752.
Using the asterisk (*) for position (p) eliminates the need to calculate positions
for consecutive fields and to change position values if you insert fields later.

13–10 Syncsort MFX Programmer’s Guide

 • p can be an equal sign (=) which indicates that the previous position should be
assigned to p. If the previous position has not yet been determined, an error
message will be generated.
When p is an equal sign (=) and there is no length specified, you can also specify
+ or – n, where n is a number from 1 through 32752.
Note that use of = for p can result in incorrect position values if fields are
inserted at some later date.
The value of the next position and the previous position can also be modified by a
POSITION operator statement. See “Using POSITION in Operator Statements” on
page 13-15.
The dictionary table, which can be printed on request (see the SYMNOUT DD
statement on page 13-24 ), displays the actual positions assigned to p when the
asterisk or equal sign is used for p. For example, consider the following dictionary
statements:

Payment_type,40,1,BI
@cash$,B’......00’
@check,B’......01’
@credt,B’......1.’
Teller,*,20,CH
Check#,=,4,ZD
CCard#,=,12,ZD

Figure 415. Sample Dictionary Statements

MFX will print the following dictionary table:

Payment_type,40,1,BI
@cash$,B’......00’
@check,B’......01’
@credt,B’......1.’
Teller,41,20,CH
Check#,41,4,ZD
CCard#,41,12,ZD

Figure 416. Sample Dictionary Table

Following is an example demonstrating the use of the *-n form. Here a field_name
is defined for the record length to be used in the RECORD control statement.

Syncsort MFX Programmer’s Guide 13–11

 field_name1,1,10,CH
.
.
.
.
field_namen,*,10,CH
rec_len,*-1

Figure 417. Sample Dictionary Statements

You would then use rec_len in the RECORD control statement, as follows:

RECORD TYPE=F,LENGTH=rec_len

Figure 418. Example of field_name in RECORD Control Statement

Specifying Length (l) in Field_name Dictionary Statements

Following are the rules for specifying length (l) in field_name dictionary
statements:
• l can be a number from 1 through 32752.
• l can be in the form of m.n for specifying a bit length, where m is a number from
0 through 32752 and n is a number from 0 through 7. Note that m and n cannot
both be zero.
• l can be an equal sign (=), which assigns the previous length to l. If the previous
length has not yet been determined when = is read, MFX will generate an error
message.
Note that use of = for l can result in incorrect length values if fields are inserted
at some later date.
If the dictionary table is printed, it will display the actual lengths that were
assigned when = was specified for l. For example, consider the following dictionary
statements:

Student_name,1,40,CH
Test_score_1,*,4,ZD
Test_score_2,*,=,ZD
Test_score_3,*,=,ZD

Figure 419. Sample Dictionary Statements

The dictionary table will reflect the following substitutions made by MFX:

13–12 Syncsort MFX Programmer’s Guide

 Student_name,1,40,CH
Test_score_1,41,4,ZD
Test_score_2,45,4,ZD
Test_score_3,49,4,ZD

Figure 420. Sample Dictionary Table

Specifying Format (f) in Field_name Dictionary Statements

Following are the rules for specifying format (f) in field_name dictionary
statements:
• f can be any of the following formats:
AC, AQ, ASL, AST, BI, CH, CLO, CSF, CSL, CST, CTO, D1, D2, DC1, DC2, DC3,
DE1, DE2, DE3, DT1, DT2, DT3, FI, FL, FS, LS, OL, OT, PD, PD0, SFF, SS,
TC1, TC2, TC3, TC4, TE1, TE2, TE3, TE4, TM1, TM2, TM3, TM4, TS, UFF,
Y2B, Y2C, Y2D, Y2DP, Y2P, Y2PP, Y2S, Y2T, Y2TP, Y2U, Y2UP, Y2V, Y2VP,
Y2W, Y2WP, Y2X, Y2XP, Y2Y, Y2YP, Y2Z, ZD
Formats can be specified using uppercase, lowercase, or mixed case letters.
When either p or l is specified in the bit form (m.n) and the bit number is not
zero ( n ≠ 0 ), then the only valid format is BI.
• f can be an equal sign (=) which indicates that the previous format should be
assigned to f. If the previous format has not yet been determined when an =
sign is used for f, MFX generates an error message.
Note that use of = for f can result in incorrect format values if fields are inserted
at some later date.
If the dictionary table is printed, it will display the actual formats MFX substituted
for =. For example, consider the following dictionary statements:

Student_name,1,40,CH
Test_score_1,*,4,ZD
Test_score_2,*,=,=
Test_score_3,*,=,=

Figure 421. Sample Dictionary Statements

After processing, the following substitutions will be reflected in the dictionary

table:

Syncsort MFX Programmer’s Guide 13–13

 Student_name,1,40,CH
Test_score_1,41,4,ZD
Test_score_2,45,4,ZD
Test_score_3,49,4,ZD

Figure 422. Sample Dictionary Table

Specifying a Parsed Field (%pp) in Field_name Dictionary State-

ments
Following are the rules for specifying a parsed field (%pp) in field_name dictionary
statements:
• pp can be a number from 0 to 999.
In the example below, four parsed fields are defined in the symbol dictionary.

Stock_symbol,%1
Current_price,%2
Sign_of_change,%3
Change_amount,%4

Figure 423. Sample Dictionary Table

These symbols are used in the control statement below.

INREC PARSE=(Stock_symbol=(ENDBEFR=C’,’,FIXLEN=4),
Current_price=(ENDBEFR=C’,’,FIXLEN=6),
Sign_of_change=(FIXLEN=1),
Change_amount=(ENDBEFR=C’,’,FIXLEN=5),
BUILD=(01:Stock_symbol,
07:Current_price,JFY=(SHIFT=RIGHT),
15:Sign_of_change,
16:Change_amount,JFY=(SHIFT=RIGHT))

Figure 424. Sample Control Statement

Following are the substitutions:

13–14 Syncsort MFX Programmer’s Guide

 INREC PARSE=(%01=(ENDBEFR=C’,’,FIXLEN=4),
%02=(ENDBEFR=C’,’,FIXLEN=6),
%03=(FIXLEN=1),
%04=(ENDBEFR=C’,’,FIXLEN=5),
BUILD=(01:%01,
07:%02,JFY=(SHIFT=RIGHT),
15:%03,
16:%04,JFY=(SHIFT=RIGHT))

Figure 425. Sample Substitutions

The Operator Statement: Rules and Syntax

The operator statement controls the position of fields you are defining.
Syntax rules for the operator statement are the same as for the other types of
dictionary statements. In addition, the operator must be specified in all uppercase
letters.
The format of the operator statement is illustrated below.

 
  ,r 
 POSITION  
  ,fieldname  
 
 SKIP,n 
 
   [ comment ]
  ,B  
  ,D  
 ALIGN   
  ,F  
   
  ,H  
 

Figure 426. Operator Statement Format

Using POSITION in Operator Statements

The POSITION parameter specifies a starting position.
• r sets the next position and the previous position to a value. The next position is
used when an asterisk (*) replaces p in a field_name statement. The previous
position is used when an equal sign (=) replaces p in a field_name statement.
Following a POSITION,r statement, either an asterisk (*) or an equal sign (=)
can be used for position (p) in the next field_name statement.
r can be any number from 1 through 32752.

Syncsort MFX Programmer’s Guide 13–15

 r can be in the form of m.n for specifying a bit position, with m being a number
from 1 through 32752 and n being a number from 0 through 7.
Following is an example where POSITION,r is used with other dictionary
statements:

POSITION,45
Volser,*,8,CH

Figure 427. Example of POSITION,r in Dictionary Statements

If the dictionary table is printed, it will reflect these substitutions:

Volser,45,8,CH

Figure 428. Sample Dictionary Table

• field_name sets the next position and the previous position to the position of the
specified field name. The next position is used when an asterisk (*) replaces p in
a field_name statement. The previous position is used when an equal sign (=)
replaces p in a field_name statement. Following a POSITION,field_name
statement, either an * or an = can be used for p in the next field_name
statement.
The field_name used with the POSITION operator can be any previously
defined field_name. As a result, POSITION,field_name allows you to map
different fields over the same locations.
Following is an example where POSITION,field_name is used with other
dictionary statements:

Filename,1,8,CH
Filetype,*,8,CH
Filemode,*,2,CH
POSITION,Filename
Filespec,*,18,=

Figure 429. Example of POSITION,field_name in Dictionary Statements

MFX will print the following dictionary table:

Filename,1,8,CH
Filetype,9,8,CH
Filemode,17,2,CH
Filespec,1,18,CH

Figure 430. Sample Dictionary Table

13–16 Syncsort MFX Programmer’s Guide

Using SKIP in Operator Statements
The SKIP parameter is used to skip unwanted positions.
• n can be any number from 1 through 32752.
• n can be in the form of u.v for specifying a bit length, where u is a number from
0 through 32752 and v is a number from 0 through 7. Note that u and v cannot
both be zero.
SKIP,n increases the next position by n bytes. The next position is used when
an asterisk (*) replaces p in a field_name statement.
Following is an example which uses SKIP,n with other dictionary statements:

Make,1,10,CH
SKIP,4
pp1,= (Last position not changed by SKIP)
Model,*,10,CH

Figure 431. Example of SKIP,n in Dictionary Statements

MFX will print the following dictionary table:

Make,1,10,CH
pp1,1
Model,15,10,CH

Figure 432. Sample Dictionary Table

Using ALIGN in Operator Statements

The ALIGN operator is used to align fields on boundaries.
• B aligns the next position on a byte boundary, for example 1, 2, 3,... The next
position is used when an asterisk (*) replaces p in a field_name statement.
Uppercase (B) and lowercase (b) are both permissible.
The following example uses ALIGN,B with other dictionary statements:

ZD1_zone,5.0,0.4,BI
ALIGN,B
Ten_bits,*,1.2,BI
ALIGN,B
nxt_byte,*,2,CH

Figure 433. Example of ALIGN,B in Dictionary Statements

MFX will print the following dictionary table:

Syncsort MFX Programmer’s Guide 13–17

 ZD1_zone,5,0.4,BI
Ten_bits,6,1.2,BI
nxt_byte,8,2,CH

Figure 434. Sample Dictionary Table

• D aligns the next position on a doubleword boundary, for example 1, 9, 17,...

The next position is used when an asterisk (*) replaces p in a field definition
statement. Uppercase (D) and lowercase (d) are both permissible.
The following example uses ALIGN,D with other dictionary statements:

Account#,42,8,CH
ALIGN,D
Balance,*,8,CSL

Figure 435. Example of ALIGN,D in Dictionary Statements

MFX will print the following dictionary table:

Account#,42,8,CH
Balance,57,8,CSL

Figure 436. Sample Dictionary Table

• F aligns the next position on a fullword boundary, for example 1, 5, 9,... The
next position is used when an asterisk (*) replaces p in a field_name statement.
Uppercase (F) and lowercase (f) are both acceptable.
The following example uses ALIGN,F with other dictionary statements:

DIVISION,34,3,FI
ALIGN,F (already aligned)
DEPT_1,*,3,FI
ALIGN,F
DEPT_2,*,3,FI

Figure 437. Example of ALIGN,F in Dictionary Statements

MFX will print the following dictionary table:

DIVISION,34,3,FI
DEPT_1,37,3,FI
DEPT_2,41,3,FI

Figure 438. Sample Dictionary Table

13–18 Syncsort MFX Programmer’s Guide

 • H aligns the next position on a halfword boundary, for example 1, 3, 5,…The
next position is used when an asterisk (*) replaces p in a field_name statement.
Uppercase (H) and lowercase (h) are both permissible.
The following example uses ALIGN,H with other dictionary statements:

box_1,1,1,BI
ALIGN,H
box_2,*,1,BI
ALIGN,H (Last position not changed by ALIGN)
sel_2,=,1,BI

Figure 439. Example of ALIGN,H in Dictionary Statements

MFX will print the following dictionary table:

box_1,1,1,BI
box_2,3,1,BI
sel_2,3,1,BI

Figure 440. Sample Dictionary Table

Activating the Dictionary Feature

Activate the MFX dictionary feature by specifying a SYMNAMES DD statement
which defines your dictionary statement data sets.
The dictionary statements can be contained in a physical sequential data set, a
member of a PDS, or DD *. Each data set must consist of fixed-length, 80-byte
records.
You may specify multiple dictionaries by concatenating data sets with the
SYMNAMES DD.
The following sample application illustrates the use of the SYMNAMES DD * .

Syncsort MFX Programmer’s Guide 13–19

 //S1 EXEC PGM=SORT
//SYMNAMES DD *
* Start of the Dictionary statements
Field1,1,10,CH First field in data record
Field2,11,4,ZD Second field in data record
.
.
.
Field40,251,15,CH Last field in data record
* End of the Dictionary statements
/*
//SYSIN DD *
SORT FIELDS=(Field2,A,Field40,D)
END
/*

Figure 441. Sample Application with SYMNAMES DD *

Using Dictionary_names in MFX Control Statements

Constant_names can be used in the following control statements:
INCLUDE
INREC
JOINKEYS
MERGE
OMIT
OUTFIL
OUTREC
RECORD
SORT
You can use a constant_name wherever you would specify a constant (X'nn...',
B'bbbb,...',C'ccc...', Y'xx...'), that is:
• A constant compared to a field in a data record (INCLUDE/OMIT)
• A constant inserted into a data record (INREC/OUTREC, OUTFIL)
• A constant used in a HEADERn/TRAILERn parameter, where n=1, 2, or 3
(OUTFIL)
You can use unsigned decimal constant_names as values for most keyword
operands in control statements, such as the LENGTH operand that is used in the
RECORD, INREC, OUTREC and OUTFIL control statements. For example, you
can specify:

13–20 Syncsort MFX Programmer’s Guide

 RECORD TYPE=V,LENGTH=reclen
where the constant_name is defined as
reclen,500
Unsigned decimal constant_names can be used for the following operands:
ABSPOS, ACCEPT, ADDPOS, DO, ENDPOS, ENDREC, FIXLEN, ID, IFOUTLEN,
INCR, LENGTH, LINES, MAXLEN, RECORDS, REPEAT, SAMPLE, SEQ,
SKIPREC, SPLIT1R, SPLITBY, START, STARTPOS, STARTREC, STOPAFT and
SUBPOS.
Note that you cannot use a dictionary_name for any of the following:
• EDIT parameter
• DATE parameter
Field_names can be used in the following MFX control statements:
DUPKEYS
INCLUDE
INREC
JOINKEYS
MERGE
OMIT
OUTFIL
OUTREC
RECORD
REFORMAT
SORT
SUM
You can use a field_name wherever you would specify a position, length, and format
(p,l,f or p,l or p), a parsed field (%pp) or an output column.

Notes on Format Substitution

When substituting field_names in control statements, MFX checks the context in
which each field_name appears and determines which field specification is
appropriate: p,l,f or only p,l. A field_name substitution results in position and
length (p,l) under the following conditions:
• There is a separate FORMAT=parameter present in the control statement.
• A format is explicitly specified for the field; that is, the field_name is followed
by a format specification.
• The field_name appears in HEADERn/TRAILERn (where n=1, 2, or 3) in an
OUTFIL statement but outside a TOTAL, TOT, SUBTOTAL, SUB, MIN,
SUBMIN, MAX, SUBMAX, AVG, SUBAVG subparameter.
• The field_name appears in an INREC/OUTREC statement or in the OUTREC
parameter of an OUTFIL statement, with the following exceptions:

Syncsort MFX Programmer’s Guide 13–21

 • The field_name is followed by an EDIT mask (Mnn), EDIT=parameter,
SIGNS=parameter, or LENGTH=parameter.
• The field_name is enclosed in parentheses.
• The field_name is an operand of an arithmetic operation (ADD/SUB/
MUL/DIV/ MOD/MIN/MAX).
• The field_name is a Y2x or Y4x field and it is not followed by H, F, D,
HEX, or a CHANGE=parameter.
• The field_name is followed by ZDC, ZDF, PDC or PDF.
• The field_name is followed by the TO=subparameter. (See “INREC,
OUTREC, OUTFIL TO Subparameter” on page 13-23.)

Specifying Field_Names after Record Reformatting

If records are reformatted, (for example by using INREC, OUTREC or an E15 or
E35 exit), resulting in different field positions, you will need to use field_names
that correspond to the new field positions in subsequent control statements. For
example, consider the following dictionary statements:

Customer_Name,1,20,CH
Customer_Address,*,20,CH
Customer_City,*,20,CH
Customer_Zip,*,9,ZD
Customer_Acct_Bal,*,12,ZD

Figure 442. Sample Dictionary Statement

Suppose the following INREC control statement is used:

INREC FIELDS=(Customer_Name,Customer_Zip,Customer_Acct_Bal)

Figure 443. Sample INREC Control Statement

Only Customer_Name, Customer_Zip and Customer_Acct_Bal will appear in the

records after INREC processing. For subsequent control statements, your
dictionary names should come from a separate dictionary that defines the new
record layout. For example:

New_Customer_Name,1,20,CH
New_Customer_Zip,*,9,ZD
New_Customer_Acct_Bal,*,12,ZD

Figure 444. Sample Dictionary Statements

13–22 Syncsort MFX Programmer’s Guide

 If the repositioned fields are given unique names, as shown above, you can
concatenate the old and new dictionaries and use both the old and new
dictionary_names, as follows:

INREC FIELDS=(Customer_Name,Customer_Zip,Customer_Acct_Bal)
SORT FIELDS=(New_Customer_Acct_Bal,A,New_Customer_Name,A)

Figure 445. Example of Dictionary Statements in Control Statements

INREC, OUTREC, OUTFIL TO Subparameter

There is a TO= subparameter available on the INREC, OUTREC and OUTFIL
control statements. With OUTFIL, it can be specified in the OUTREC, HEADER
and TRAILER parameters. The TO= subparameter can be specified wherever there
is an fo parameter which is used to define the output numeric data format of an
expression. For example,

OUTREC FIELDS=(1,8,BI,ZD) is equivalent to

OUTREC FIELDS=(1,8,BI,TO=ZD)

In general, there is no reason to include the ‘TO=’ because the meaning is the same.
However, with field_name substitution, there can be different outcomes depending
on whether or not ‘TO=’ is used. For instance, if a field_name were defined as:

FIELD1,1,8,BI

And it was used in a substitution on an OUTREC statement as:

OUTREC FIELDS=(FIELD1,ZD)

MFX would assume the ZD should replace the BI in the context of the OUTREC
statement, resulting in the following substitution:

OUTREC FIELDS=(1,8,ZD)

If the OUTREC statement were specified as:

OUTREC FIELDS=(FIELD1,TO=ZD)

MFX would create the following substitution:

Syncsort MFX Programmer’s Guide 13–23

 OUTREC FIELDS=(1,8,BI,TO=ZD)

To ensure that there are no ambiguities with the use of data dictionary_names and
certain types of data conversions, you should use the TO= syntax when necessary.
Since the PDC, PDF, ZDC and ZDF formats can only be specified as output formats,
there is no ambiguity and the ‘TO=’ form is not required.

Specifying Dictionary Listing Output

The SYMNOUT DD statement allows you to specify a location where you would
like the dictionary listing to be printed.
The dictionary listing consists of two parts.
• The first part is the user-provided dictionary statements.
• The second part is the dictionary table generated by MFX from the dictionary
statements.
The DCB attributes for the SYMNOUT data set are: RECFM=FBA and
LRECL=121.

Error Handling for Dictionary Statements

MFX will check each dictionary statement for errors. If an error is encountered, an
error message will be generated. MFX stops scanning a dictionary statement at the
first error, and resumes with the next dictionary statement.
Once an error has been detected, positions calculated with the use of an asterisk (*)
for p or with the POSITION operator in subsequent dictionary statements will not
be validated. If an error is detected in any dictionary statement, MFX will
terminate processing after all dictionary statements are read.
If MFX detects an error in a control statement while substitution is taking place, it
may respond in either of the following ways:
• Print the statement that was in error, followed by a corresponding error
message, then continue with the next statement and terminate when all
substitutions have been completed.
• Stop the substitution for the statement in error and continue processing, letting
subsequent processing handle the error. If this occurs, the original field or

13–24 Syncsort MFX Programmer’s Guide

 constant_ name rather than the substituted value may be displayed in a
translated statement.
If there are no errors during the substitution process, MFX will substitute values
for field_names and constant_names wherever they are valid. If substituted values
prove invalid for a particular statement or parameter, this situation will be
detected after the substitution has been performed.

Sample Dictionary Statements

This section provides some examples of dictionary statements and their usage.
Example 1

Order#,10,8,CH
Catlog#,18,6,CH

Color,24,1,CH
Grey,C’G’
Red,C’R’
red,C’r’
Black,C’B’
White,C’W’

Qty,25,3,ZD
Price,28,4,PD

Figure 446. Sample Dictionary Statements

This example uses leading blanks before some color dictionary_names to create
indentation for clarity.
A blank statement is used before and after the "color" section. Such blank
statements are ignored but can be printed.
Dictionary_names are case-sensitive. Thus, Red and red are separate
dictionary_names.
Example 2

Street,38,45,CH
City,*,26,CH City is “83,26,CH”
State,*,2,CH State is “109,2,CH”
Zip,*,9,ZD Zip is “111,9,ZD”

Figure 447. Example of Use of Asterisk to Replace Position (p)

Syncsort MFX Programmer’s Guide 13–25

 In this example, note how position (p) is replaced with an asterisk (*) to indicate
that the value of p should be the next position after the previous field. This is a
powerful feature. Using an asterisk (*) for position allows you to define adjoining
fields automatically. Thus, if a field specification changes, it is not necessary to
calculate and specify the changed positions of subsequent fields.
Example 3

Name,5,40,CH
SKIP,20
Phone#,*,11,CH Phone# is “65,11,CH”

Figure 448. Example of Use of SKIP Operator

In this example, the SKIP operator advances the next position by n bytes. Since the
position of the next field is specified with an asterisk (*), the position will be
calculated automatically.
Example 4

Date_of_Birth,10,6,Y2T A 6-byte field in yymmdd format

DOB_YY,=,2,Y2C Mapping byte 1,2 of Date_of_Birth
DOB_MM,*,=,ZD Mapping byte 3,4 of Date_of_Birth
DOB_DD,*,=,= Mapping byte 5,6 of Date_of_Birth

Figure 449. Example of Use of Equal Sign to Replace p,l, or f

In this example, an equal sign (=) is used instead of p, l or f. The equal sign assigns
the previous position, length or format to the equal sign, mapping one field onto
another.
Although field_names and constant_names can usually be specified in any order,
using the asterisk (*) or equal sign (=) forces a dependency on field order.
Note the use of comments in the above examples. As for MFX control statements, a
blank after the value indicates the beginning of a comment. You can also use
comment statements, which begin with an asterisk (*) in column 1.
Example 5

//SYMNAMES DD DSN=PARTS.LAYOUT,DISP=SHR
// DD DSN=MFC.LAYOUT,DISP=SHR
// DD DSN=STOCK.LAYOUT,DISP=SHR

Figure 450. Sample SYMNAMES DD Statement

13–26 Syncsort MFX Programmer’s Guide

 This example uses concatenated dictionaries. Suppose the dictionaries contain
field_name statements as follows:

The PARTS.LAYOUT dictionary contains:

Part#,1,8,CH
Desc$,*,20,CH

The MFC.LAYOUT dictionary contains:

Mfc_code,*,6,CH
Mfc_part#,*,10,CH

The STOCK.LAYOUT dictionary contains:

Stock_Shelf#,*,4,BI
Stock_Bin#,*,4,BI
Stock_Qty,*,4,PD

Figure 451. Sample field_name Statements in Dictionaries

MFX will print the following dictionary table:

Part#,1,8,CH
Desc$,9,20,CH
Mfc_code,29,6,CH
Mfc_part#,35,10,CH
Stock_Shelf#,45,4,BI
Stock_Bin#,49,4,BI
Stock_Qty,53,4,PD

Figure 452. Sample Dictionary Table

Field_names from all three dictionaries in the above example could be used in MFX
control statements as follows:

OUTFIL FNAME=ORDER,INCLUDE=(Stock_Qty,LT,100),
OUTREC=(Part#,10X,Mfc_code,5X,Mfc_part#)
OUTFIL FNAME=INTEL,INCLUDE=(Mfc_code,EQ,C’INTEL’),
OUTREC=(Part#,10X,Stock_Shelf#,LENGTH=6,
8X,Stock_Bin#,LENGTH=6,
8X,Stock_Qty,LENGTH=8)

Figure 453. Sample field_names in Control Statements

Syncsort MFX Programmer’s Guide 13–27

 Example 6
This example demonstrates some circumstances where format specifications are
used and are not used in control statements.
In your symbols dictionary you can specify a field_name with p, l, and f, then use
the field_name in a control statement that only requires p and l. MFX will
substitute p and l during processing and ignore the f specification under certain
circumstances. For example, consider the following dictionary statements:

Account#,1,12,CH
CheckNumber,13,4,ZD
CheckDate,17,6,Y2T
CheckAmount,23,6,PD

Figure 454. Example of p,l,f in Dictionary Statements

Suppose you use these field_names in the following control statements:

SORT FIELDS=(Account#,A,CheckDate,A)
DUPKEYS SUM=(CheckAmount),FORMAT=PD
OUTFIL OUTREC=(Account#,CheckDate,CheckAmount,M23)

Figure 455. Example of field_names in Control Statements

Based on the field_names defined in the dictionary statements, MFX will make the
following substitutions:
• In the SORT statement, 'Account#' is replaced with '1,12,CH' and 'CheckDate' is
replaced with '17,6,Y2T'.
• In the DUPKEYS statement, 'CheckAmount' is replaced with '23,6'.
• In the OUTFIL statement, 'Account#' is replaced with '1,12', 'CheckDate' is
replaced with '17,6,Y2T' because of the special nature of the Y2T format, and
'CheckAmount' is replaced with '23,6,PD' because it is followed by an EDIT
mask.
Here are the resulting control statements: :

SORT FIELDS=(1,12,CH,A,17,6,Y2T,A)
DUPKEYS SUM=(23,6),FORMAT=PD
OUTFIL OUTREC=(1,12,17,6,Y2T,23,6,PD,M23)

Figure 456. Example of p,l in Control Statements

13–28 Syncsort MFX Programmer’s Guide

Using JCL SET and PROC Symbols to Create
Dictionary_Names
There is another way to establish dictionary_names directly in JCL statements
rather than in a SYMNAMES data set. For JCL-initiated MFX applications only,
the data from JCL SET and PROC symbols can be used together with the MFX
JPn PARM options to create character string dictionary_names. Text strings and
system symbols can also be used. These JPn dictionary_names can then be used in
MFX control statements in the same manner as the SYMNAMES
dictionary_names. The ability to alter JCL to dynamically change control
statements that are often contained in data sets can be very useful.
On the EXEC statement, specify PARM='…,JPn''string'',…' where n is from 0 to 9.
The quotes delimit the start and end of the string, so the string should not contain
any imbedded quotes. If apostrophes are required, two should be specified for each
one so as not to terminate the PARM field. Up to 10 such JPn PARMs can be used
in one PARM field.
The string may contain any combination of
• SET or PROC symbols from the JCL. These are prefaced by an ampersand:
&symbol
• System symbols, which are also prefaced by an ampersand, such as
&JOBNAME or &YYMMDD
• Any characters except quotes; characters after a symbol name may be appended
with a period: &symbol.text
JPn dictionary_names are listed in the optional SYMNAMES DD output data set
before any other dictionary_names that have been defined in the SYMNAMES DD
data set. They will be built as

JPn,S'string'

Figure 457. JPn format in SYMNAMES DD

Syncsort MFX Programmer’s Guide 13–29

 Example 1 Using JCL SET symbols
Select data for only certain states where the list of states will vary:

// SET STATE1='NY'
// SET STATE2='NJ'
// SET STATE3='PA'
// SET STATE4='CT'
// SET STATE5=' '
// SET STATE6=' '
// SET STATE7=' '
// SET STATE8=' '
//SELECT1 EXEC PGM=SORT,
// PARM=('JP1"&STATE1",JP2"&STATE2",JP3"&STATE3",JP4"&STATE4"',
// 'JP5"&STATE5",JP6"&STATE6",JP7"&STATE7",JP8"&STATE8"')
//SYSOUT DD SYSOUT=*
//SYMNOUT DD SYSOUT=*
//SORTIN DD *
TALLAHASSEE FL
TRENTON NJ
TOPEKA KS
HARTFORD CT
SACRAMENTO CA
//SORTOUT DD SYSOUT=*
//SYSIN DD *
INCLUDE COND=(15,2,CH,EQ,L(JP1,JP2,JP3,JP4,JP5,JP6,JP7,JP8))
SORT FIELDS=(1,16,CH,A)

Figure 458. Example of using JCL SET symbols

After data dictionary symbol substitution, the INCLUDE statement becomes

INCLUDE COND=(15,2,CH,EQ,L(C'NY',C'NJ',C'PA',C'CT',C' ',C' ',

C' ',C' '))

Figure 459. Using JCL SET symbols

The output is

HARTFORD CT
TRENTON NJ

Figure 460. Using JCL SET symbols

13–30 Syncsort MFX Programmer’s Guide

 Example 2 Using system symbols
Select data for the current month using system symbols for MM and YYYY:

//SELECT2 EXEC PGM=SORT,PARM='JP1"&MON&YR4"'

//SYSOUT DD SYSOUT=*
//SYMNOUT DD SYSOUT=*
//SORTIN DD *
062013
072012
072013
012013
//SORTOUT DD SYSOUT=*
//SYSIN DD *
INCLUDE COND=(1,6,CH,EQ,JP1) SELECT DATA FOR THIS MONTH ONLY
SORT FIELDS=(1,6,CH,A)

Figure 461. Example of using system symbols

Assuming this JCL was executed in July 2013, after data dictionary symbol
substitution the INCLUDE statement becomes

INCLUDE COND=(1,6,CH,EQ,C'072013')

Figure 462. Using system symbols

The output would be

072013

Figure 463. Using system symbols

Syncsort MFX Programmer’s Guide 13–31

 Example 3 Using PROC symbols and text
Produce a report for one particular month, which is selectable from the EXEC
statement for a PROC:

//RPRTPROC PROC MONTH=,YEAR

//MONREPRT EXEC PGM=SORT,
// PARM='JP1"&MONTH",JP2"&YEAR",JP3"REPORT FOR &MONTH &YEAR"'
//SYSOUT DD SYSOUT=*
//SYMNOUT DD SYSOUT=*
//RPRTPROC PEND
//********
//REPORT EXEC RPRTPROC,MONTH=11,YEAR=2013
//SORTIN DD *
12252013 11111
11222013 16797
07042013 07446
11012013 33458
11012013 20142
12312012 09876
04012013 54321
11282013 66666
//SORTOUT DD SYSOUT=*
//SYSIN DD *
OUTFIL HEADER1(JP3),INCLUDE=(1,2,CH,EQ,JP1,AND,5,4,CH,EQ,JP2)
SORT FIELDS=(1,14,CH,A)

Figure 464. Example of using PROC symbols and text

After data dictionary symbol substitution, the OUTFIL statement becomes

OUTFIL HEADER1(C'REPORT FOR 11 2013'),

INCLUDE=(1,2,CH,EQ,C'11',AND,5,4,CH,EQ,C'2013')

Figure 465. Using PROC symbols and text

The report is produced for the selected month only:

REPORT FOR 11 2013

11012013 20142
11012013 33458
11222013 16797
11282013 66666

Figure 466. Using PROC symbols and text

13–32 Syncsort MFX Programmer’s Guide

Chapter 14 Performance
Considerations

Disk Sort? MAXSORT? PARASORT?

Disk Sort provides the current, established sorting technique, suitable for most
sort/merge applications. Intermediate storage is allocated on disk devices and the
sort size is limited by the allocated disk space plus secondary extents automatically
obtained by the sort.
MAXSORT, MFX’s maximum capacity sorting technique, is not limited by disk
space availability. MAXSORT determines how much data can be sorted using the
available disk work space and divides SORTIN into SORTWK-manageable
segments; the sorted segments are stored on tape or disk for a later, automatic
merge. MAXSORT makes all the Disk Sort operational optimizing features and
modern programming options available to large sorts, and additionally provides an
enhanced breakpoint/restart capability for greater scheduling flexibility--you can
stop MAXSORT processing at selected intervals without loss of sorted output.
PARASORT improves elapsed time performance for sorts whose input is read from
a multi-volume tape data set and/or concatenated tape data sets. The performance
improvement from PARASORT is a result of processing the SORTIN input volumes
in a parallel fashion. PARASORT requires two to eight times the current number of
tape units, depending upon resource availability and the degree of improvement
desired. PARASORT automatically manages the tape units and minimizes the use
of the tape drive resources by deallocating excess tape drives during initialization
and releasing all the extra units at the end of the sort input phase.

Syncsort MFX Programmer’s Guide 14–1

JCL Sorts vs. Program-Invoked Sorts
When MFX is initiated from a COBOL program, the calling program handles I/O,
remains in storage, and generally retards sort execution. MFX will yield maximum
performance through proper synchronization of all data whenever it has control of
the sorting process, i.e., whenever // EXEC PGM= SYNCSORT is used.
From the point of view of performance, the JCL-initiated sort execution has the
advantage. Whenever possible, tasks incidental to the sort/merge/copy process
should be handled via MFX control statements. Where this is not possible, the JCL/
control stream should be supplemented with user-written exit routines. Ideally, the
exit routines exist as load modules, so that they do not require link-editing every
time the job is run. MFX permits exit routines to be written in COBOL, C,
FORTRAN, REXX, or Assembler language.
If you must invoke the sort from a COBOL program, you may improve sort
performance by passing an accurate FILSZ=n/En parameter via $ORTPARM.

Control Statement Issues

MFX control statements can be used to eliminate records from the input file
(INCLUDE/OMIT), obtain the minimum/maximum/sum/average of fields in equal-
keyed records, and/or eliminate equal-keyed records (SUM or DUPKEYS), reformat
records (INREC/OUTREC), set up multiple output files (OUTFIL) or write
formatted reports (OUTFIL statement with HEADER, TRAILER, SECTIONS and
OUTFIL parameters. These control statements provide a high performance
alternative to the use of exits and invoking programs. The tasks they address are
those which are most frequently executed and/or improve sort performance. Since
sort throughput is in part a function of the number of bytes that are to be
manipulated, considerable performance savings can result from using the
INCLUDE/OMIT statement to eliminate irrelevant records; INCLUDE/OMIT
affects the data set prior to sorting/merging/copying.
The SKIPREC and STOPAFT parameters are recommended for test runs of sorting
applications for the same reason. When the file bias is high enough for a significant
number of records to be summed early in the sort, SUM or DUPKEYS will also
provide performance gains if the XSUM or XDUP option has not also been selected.
When reformatting records, it is desirable to minimize the amount of data that
must pass through the sort process. Other things being equal, INREC should be
used to shorten records, OUTREC to lengthen them.

14–2 Syncsort MFX Programmer’s Guide

The Efficient Use of PARMs
There are four programming PARMs that may have a significant effect on sort
performance: CMP, EQUALS, STOPAFT, and SKIPREC.
The CMP PARM specifies the kind of compare operation to be used for sort/merge
control fields up to 16 bytes long, bearing the format code PD or ZD. When
CMP=CPD, the default, is used, ZD fields are PACK’ed and then compared. Invalid
PD data may cause a system 0C7 abend and program termination. The integrity of
fields labelled “ZD” is only guaranteed when they contain valid ZD data. The
delivered default of the VLTEST PARM supports CMP=CPD, as do certain other
VLTEST PARM values. Whenever possible, set CMP= CPD for better sort
performance.
The alternative, CMP=CLC, is a more costly option--it forces the sort to extract
potentially invalid PD and ZD fields and do a certain amount of data manipulation
to obtain valid sign comparisons.
The EQUALS PARM instructs the sort/merge to preserve the order of equal-keyed
records. EQUALS will have a slight but generally significant impact on sort
performance. By making EQUALS available on an individual sort basis, MFX
makes this programming option available where it is needed, without imposing it
on the installation’s more routine jobs. For sort efficiency, use EQUALS only where
the preservation of the input order of equal-keyed records is important.
Users interested in sort performance should specify the STOPAFT PARM in test
runs of the sort. With STOPAFT=n, only the first n records of the input file will be
sorted. By reducing the number of records to be processed, STOPAFT improves sort
performance. If additional tests are necessary, the SKIPREC PARM can be used
together with STOPAFT to select a different subset of the SORTIN data set.

Optimizing System Resources

The efficiency of sort processing is measured in terms of the performance measures
of CPU time, elapsed time, and I/O activity. Ordinarily, when MFX performs a sort,
it seeks to balance these performance measures in a way that yields the best
overall sort performance. It is possible, however, to define a particular performance
measure as more important than others for a particular job.
This can be done through MFX’s Dynamic Storage Management (DSM) facility,
which makes available four optimization modes for sort processing. These are
BALANCE, CPU, ELAP and IO. BALANCE is the default optimization mode which
provides the best overall balance between CPU time, sort elapsed time and I/O
activity to SORTIN, SORTOUT and SORTWK.

Syncsort MFX Programmer’s Guide 14–3

 If CPU time is given the highest priority, MFX will minimize this resource at the
expense of elapsed time and I/O activity. Selecting ELAP as the optimization mode
will cause MFX to minimize the elapsed (wall clock) time of each sort, usually at
some expense of the sort’s CPU time. Likewise, if IO is selected as the optimization
mode, MFX will minimize the I/O activity (EXCPs) performed by the sorts.

The Incore Sort

Whenever there is sufficient memory, the standard Disk Sort may sort all the input
data within its memory area, without writing to any of the work data sets that may
have been provided. Sufficient memory, as discussed here, means that MFX’s
memory area/address space is large enough to hold the MFX program, all of the
input data, SORTIN or SORTOUT buffers (whichever are larger) and, if work data
sets are allocated, SORTWKxx buffers.
The incore sort is not available to Disk Sorts taking checkpoints, using SUM,
DUPKEYS, OUTREC, OUTFIL, an E14 or E16 exit routine, or producing VSAM
output.

Disk Space Considerations

Tuning Disk Space Allocations
With the operational feature RELEASE turned ON (this is the delivered default),
MFX automatically supplements and releases any disk space you have allocated for
intermediate storage, making the allocation of the correct amount of SORTWK
space an automatic, sort-controlled process. For general sorting purposes, you don’t
need to be concerned with precise SORTWK space allocations. However, allocating
SORTWK space in cylinders, rather than blocks or tracks, will usually yield
optimal performance.
For best performance with file sizes greater than 30 megabytes, especially when
DYNALLOC is not enabled, allocate the required space across 4 to 6 SORTWK
devices.
Message WER124I is provided in some applications in order to permit those who
are interested in a finely-tuned sort execution to improve intermediate storage
allocation for future runs. Routinely overallocating SORTWK, relying on
RELEASE=ON, will delay sort step execution until all the space requested
(including the excess space) is available, and will waste this excess space until its
released at the end of Phase 1. Routinely underallocating by a large amount
assumes that the needed storage will always be physically available. If, for some

14–4 Syncsort MFX Programmer’s Guide

 reason, the required storage cannot be obtained on any volume assigned for sort
work areas, MFX will terminate with a SORT CAPACITY EXCEEDED error.
A sort is considered to finely tuned when WER124I reports an overallocation factor
between 1.00 and 1.50.

The Impact of Disk Space on the Work Data Sets on MFX

MFX’s work data set disk space management is automated to a very high degree. It
can:
• Automatically correct the underallocation of disk space by obtaining secondary
allocations of disk space, as needed. This prevents costly SORT CAPACITY
EXCEEDED terminations.
• Automatically release excess disk space at the completion of Phase 1. The space
immediately becomes available for allocations to other jobs.
• Dynamically allocate work data sets through the z/OS DYNALLOC capability.
You can improve the efficiency of disk space usage by allocating optimally at the
outset.

Disk Sort Intermediate Storage Calculation Formulas

Approximate Number of Tracks Required = A X B X 1.3
C
where:
A = Number of records to be sorted.
B = Average record length of data being sorted.
C = Track capacity of the work device:

DEVICE TYPE TRACK CAPACITY IN BYTES

3380 47,476

3390 56,664

9345 46,456

Table 68. Device Type and Track Capacity

Allocating disk storage space in cylinders rather than tracks will improve the
performance of MFX. When converting tracks to cylinders, round the number of
cylinders up to the higher number. For example, if 9.5 cylinders are needed,
allocate 10 cylinders.

Special Considerations Concerning MFX’s Disk Space Management

on Work Data Sets
MFX implements the automatic space management facility by reading the JFCB
and modifying the SPACE parameter to enter a secondary allocation quantity (or

Syncsort MFX Programmer’s Guide 14–5

 accept one that was coded) and the RLSE subparameter. The JFCB is the z/OS
control block that represents the DD statement.
Several MFX options may be implemented which affect the disk space management
of the sort. (See the Default Options chapter of the Installation Guide.) Specific
functions of these features are as follows.

1. The automatic release of excess SORTWK space can be selectively suppressed.

2. The amount of secondary space per allocation can be modified.

3. The use of RLSE can be suppressed.

4. The use of space release is suppressed for small sorts, including the incore sort, in order
to minimize system overhead. For non-incore sorts, if the file size is less than 4
megabytes, space release is normally suppressed. The 4 megabyte threshold may be
altered.

5. The use of space release is normally suppressed for all invoked sorts to prevent SORT
CAPACITY EXCEEDED termination when MFX is invoked more than once by a single
program. If the first sort involves a modest volume of data, and causes space release to
make the work data sets smaller, and the second sort is larger, the second sort might
not find sufficient work space. Your installation can turn on space release for invoked
sorts and thus save disk space. This very rarely causes problems because (1) few
programs invoke the sort more than once and (2) MFX’s automatic secondary allocation
normally prevents a SORT CAPACITY EXCEEDED termination.

6. MFX can routinely DYNALLOC data sets for every run.

Other factors which may result in suppression of these features include:

1. Automatic release is suppressed for permanent data sets unless additional sort work
space has been allocated.

2. Automatic release is normally suppressed by the installation for initiator-dedicated

data sets.

MFX uses normal z/OS facilities to obtain secondary allocations on work data sets.
Consequently, the sort, like any other program under z/OS, is restricted to sixteen
extents per data set. MFX, however, will recover from system B37 ABENDS that
other programs might encounter in attempting secondary allocations. If MFX
determines that a particular sort work data set cannot sustain a secondary
allocation because it already has sixteen extents or because there is not enough
space left on the volume, it does not attempt secondary allocation on that data set.
MFX further checks all other work data sets, and if none of them can sustain a
secondary allocation, it must abort with SORT CAPACITY EXCEEDED.
MFX often avoids the use of one or more work data sets to minimize overall system
conflict, as, for instance, between SORTIN and a work data set. It may obtain
secondary allocation on some data sets while releasing on others.

14–6 Syncsort MFX Programmer’s Guide

The Coding and Use of Checkpoint-Restart
Occasionally, a hardware failure may prevent the successful completion of a sort or
merge. Examples include a physically defective output volume or device, or a
failure of the operating system for reasons unrelated to the sort. Since sorts tend to
consume more system resources than any other type of application, it may be
advantageous to be able to resume execution just before the failure occurred rather
than restart the job at the beginning of the failed job step. MFX provides this
restart capability through its support of the standard z/OS Checkpoint-Restart
feature.
To instruct MFX to take checkpoints, code CKPT or CHKPT (either spelling) on the
SORT/MERGE control statement and supply a SORTCKPT DD statement.
For a sort, checkpoints are taken at the beginning of Phase 3 before the output data
sets (if any) are opened, and at every end-of-volume of a SORTOUT data set when
OUTFIL is not in use. An operator may then restart the sort at Phase 3 or at any
end-of-volume checkpoint. If necessary, a new output volume or device with
identical characteristics to the defective volume or device may be substituted.
For a merge or copy, MFX takes a checkpoint at every end-of-volume of a
SORTOUT data set when OUTFIL is not in use.
Checkpoints cannot be taken within a user exit routine.

The DISP Parameter for SORTCKPT, SORTWKxx and SORTOUT

Data Sets
The coding of the DISP parameter for these data sets depends in part on the
PARM-specified response to an unsuccessful sort. There are four cases:
• NOIOERR and NORC16 (These are the delivered defaults.)
• IOERR=ABE and RC16=ABE
• IOERR=ABE and NORC16
• NOIOERR and RC16=ABE
When return code 16 is issued by the unsuccessful sort (i.e., when an I/O error
occurs and NOIOERR is set or, for other errors, when NORC16 is set), the second
subparameter of the DISP parameter should be specified as KEEP or CATLG.
When the unsuccessful sort causes a user abend (i.e., when IOERR=ABE for I/O
errors, RC16=ABE for other errors), the third subparameter of the DISP parameter
should be specified as KEEP or CATLG. Thus, with NOIOERR and RC16=ABE or
with IOERR=ABE and NORC16, both the second and the third DISP subparameter
should be specified as KEEP or CATLG. Unless the DISP parameter is coded in
accordance with these two PARM values, restart will be impossible.
It is recommended that these data sets be deleted upon successful completion of the
sort. This can be done by coding the COND parameter for an IEFBR14 step to

Syncsort MFX Programmer’s Guide 14–7

 follow the sort step in the jobstream. The COND parameter makes the IEFBR14
(data set deletion) execution depend upon the successful completion of the previous
step (the sort).

The SORTCKPT Data Set

Assign a permanent DSN to the SORTCKPT DD statement and specify the UNIT,
SPACE and VOL=SER parameters to make the operator’s job easier should a
deferred restart become necessary.

//SORTCKPT DD UNIT=3390,DSN=SORT.CKPT,
// SPACE=(CYL,(1,1)),
// VOL=SER=WORK01,DISP=(MOD,KEEP,KEEP)

Figure 467. Sample SORTCKPT DD Statement

The SORTWKxx Data Set(s)

Assign a permanent DSN to every SORTWKxx DD statement and specify the
UNIT, SPACE and VOL=SER parameters in case a deferred restart becomes
necessary. Avoid using passed data sets, JCL refer-backs, and any other references
which would make the JCL following the restart dependent on the JCL preceding
the restart.
Note that the SORTCKPT data set and the SORTWKxx data set(s) may reside on
the same direct access device without loss of efficiency.

//SORTWK01 DD UNIT=3390,DSN=SORT.WK01,
// SPACE=(CYL,(20,10)),
// VOL=SER=WORK02,DISP=(,KEEP,KEEP)

Figure 468. Sample SORTCKPT DD Statement

Automatic Checkpoint-Restart
With automatic checkpoint-restart, the operating system will ask the operator
whether an unsuccessful/abending step should be restarted. A “yes” reply instructs
the system to restart the job at the last checkpoint taken. If the operator replies
“no,” the job will still be eligible for deferred checkpoint-restart, but its control
statements will have to be modified before the job is resubmitted.
The requirements for automatic checkpoint-restart are:
• The sort step must have a unique name.
• The JOB statement must specify RD=R and MSGLEVEL=1.

14–8 Syncsort MFX Programmer’s Guide

 • All system completion codes with which the sort may abend should be defined
at system generation time as being eligible for restart. If the RC16=ABE and/or
IOERR=ABE options are in effect, for example, then user abend codes 16 and/or
999 must be eligible for restart.
• User-written exit routines and calling programs may not issue the STIMER
macro.

//AUTOCKPT JOB (1101,2333),P.ARBEAU,RD=R,

// MSGLEVEL=(1,1)
//XVISORT EXEC PGM=SORT,
// PARM='RC16=ABE,IOERR=ABE'
//SYSOUT DD SYSOUT=A
//SORTIN DD DSN=XVI.SORTIN,DISP=OLD
//SORTOUT DD UNIT=TAPE,DISP=(,CATLG,KEEP),
// DSN=XVI.SORTOUT
//SORTWK01 DD UNIT=3390,DISP=(,DELETE,KEEP),
// VOL=SER=WORK01,DSN=XVI.SORTWK01,
// SPACE=(CYL,40)
//SORTWK02 DD UNIT=3390,DISP=(,DELETE,KEEP),
// VOL=SER=WORK02,DSN=XIV.SORTWK02,
// SPACE=(CYL,40)
//SORTCKPT DD UNIT=3390,DISP=(,DELETE,KEEP),
// VOL=SER=WORK02,DSN=XVI.SORTCKPT,
// SPACE=(CYL,(1,1))
//SYSIN DD *
SORT FIELDS=(1,10,CH,A),CKPT
/*

Figure 469. Sample Automatic Checkpoint-Restart JCL Stream

Deferred Checkpoint-Restart
Unlike automatic checkpoint-restart, deferred checkpoint-restart requires that
certain JCL changes be made before resubmitting the job.
The requirements for a deferred restart are:
• A SYSCHK DD statement must appear immediately before the first EXEC
statement in the job. The SYSCHK DD must use the same DSN name as the
SORTCKPT DD of the sort that failed. Specify UNIT, VOL=SER, and
DISP=(OLD,KEEP).
• The RESTART parameter must be specified, and must provide the job
stepname and the PROC stepname (if any) associated with the step containing
the failed sort, as the first subparameter. (Separate the two stepnames by a
period.) The second subparameter should contain the checkpoint ID of the last
checkpoint taken before the sort failed. This can be determined from the console

Syncsort MFX Programmer’s Guide 14–9

 messages given for the job. For JCL sorts, the ID is usually "Cnnnnnnn,"
referring to the sequence number assigned by the operating system.
• SORTIN and SYSIN DD DUMMY statements are permissible if the program is
being restarted at a point where they are no longer needed.

//DEFCKPT JOB (5433,2333),PAT.TAIG.NANT, 1

// RD=R,MSGLEVEL=(1,1),
// RESTART=(XVISORT,C0000001) 1
//SYSCHK DD UNIT=3390,DISP=(OLD,KEEP), 1
// VOL=SER=WORK02,
// DSN=XVI.SORTCKPT 1
//XVISORT EXEC PGM=SORT,
// PARM='RC16=ABE,IOERR=ABE'
//SYSOUT DD SYSOUT=A
//SORTIN DD DUMMY 1
//SORTOUT DD UNIT=TAPE,DISP=(,CATLG,KEEP),
// DSN=XVI.SORTOUT
//SORTWK01 DD UNIT=3390,DISP=(OLD,DELETE,KEEP),
// VOL=SER=WORK01,DSN=XVI.SORTWK01,
// SPACE=(CYL,40)
//SORTWK02 DD UNIT=3390,DISP=(OLD,DELETE,KEEP),
// VOL=SER=WORK02,DSN=SVI.SORTWK02,
// SPACE=(CYL,40)
//SORTCKPT DD UNIT=3390,DISP=(MOD,DELETE,KEEP),
// VOL=SER=WORK02,DSN=XVI.SORTCKPT,
// SPACE=(CYL,(1,1))
//SYSIN DD DUMMY 1

Figure 470. Sample Deferred Checkpoint-Restart JCL Stream

1. This JCL differs from automatic checkpoint-restart JCL.

Optimizing Data Set Placement

The Impact of Work Devices on MFX
The performance of MFX is almost totally independent of the number of work data
sets. The sort’s performance may, however, be strongly influenced by the number of
devices to which the work data sets are allocated. Generally, for any sort of
significant size, the more work devices, the better the sort can perform. If the sort
file size is small, however, performance improvements might be outweighed by
increased overhead in managing the extra data sets. Increasing the number of
work devices will:

14–10 Syncsort MFX Programmer’s Guide

 1. Improve the overlap between CPU and I/O processing.

2. Improve the effectiveness of MFX’s integrated activity monitoring.

3. Reduce the likelihood of SORT CAPACITY EXCEEDED.

Increasing the number of data sets without increasing the number of devices will
only increase overhead.

Device Type Considerations

Avoid using a mixture of device types with different track capacities for the work
data sets, since MFX sacrifices some efficiency if this is the case.
If you must choose between two different disk device types for the work data sets,
use the faster; if they are close in speed, use the one with the larger track size.
Avoid the use of VIO data sets for work data sets.

Syncsort MFX Programmer’s Guide 14–11

14–12 Syncsort MFX Programmer’s Guide
Chapter 15 The HISTOGRM Utility
Program

What Is HISTOGRM?
HISTOGRM is a separate program which is used to gain information about
variable-length files. The program scans a variable-length file and provides
information which can then be used to run more efficient sorts. HISTOGRM can
report the:
• Block count for minimum and maximum block lengths
• Record count for minimum and maximum record lengths
• Average record length
• Total number of bytes in the file
• Total number of blocks in the file
• L6 value (average work space) for variable-length records
• L7 value (segment length) for variable-length records
HISTOGRM can be used to analyze variable-length records in a VSAM entry-
sequenced or key-sequenced data set. When HISTOGRM processes a VSAM file
only record information is gathered; block statistics are not produced.

Using HISTOGRM to Determine L6 and L7 Values for MFX

The L6 and L7 values HISTOGRM calculates are passed to MFX via the L6, L7
PARM options or the l6, l7 values in the LENGTH parameter of the RECORD
control statement. (When there is a conflict, the PARM specification takes
precedence.) These values are ignored in a merge or copy application.

© 2010, 2018 Syncsort Incorporated 15–1

Control Parameters for HISTOGRM

Control Parameters for HISTOGRM

The control parameters are outlined below; defaults are underlined. To specify
other values, include a control statement in the SYSIN DD portion of the job
control stream. Parameters may appear anywhere through column 71, provided
they are separated by commas with no intervening blanks.

NRECS

 ALL 
NRECS=  
 nnn 

Figure 471. NRECS Format

Tells how many records to scan in the variable-length file.

WIDTH

 20 
WIDTH=  
 nnnn 

Figure 472. WIDTH Format

Indicates the range between minimum and maximum block lengths and the
minimum and maximum record lengths in each group of the HISTOGRM output.
The number specified for the WIDTH value must be a multiple of 4. (4, 8, 12, . . .
See examples of block and record HISTOGRMs that follow.) Adjust this range
based on the characteristics of the file (the lengths of the shortest and longest
record) and the desired length of HISTOGRM.

DEVWK

 3380 
DEVWK=  3390 
 

Figure 473. DEVWK Format

Tells the type of disk device that will be used for intermediate storage when the
sort is run. Specify the device number if HISTOGRM is to calculate L6 and L7.

15–2 MFX for z/OS Programmer’s Guide

 Control Parameters for HISTOGRM

KEYL

 20 
KEYL=  
 nnnn 

Figure 474. KEYL Format

Gives the end location of the last control field in the record. Specify a value for
KEYL if HISTOGRM is to calculate L6 and L7.

BIGREC

 20 
 
BIGREC=  nnnn 
 
 MAX 

Figure 475. BIGREC Format

Specifies the maximum number of HIS025I messages that will be issued in a

HISTOGRM execution. When HISTOGRM processes a large file, this message may
be generated as often as once for each record in the file. BIGREC limits the number
of HIS025I messages that will be issued in each execution. HISTOGRM processing
continues, but no further messages are issued once the BIGREC value is reached.

BLOCK

 BLOCK 
 
 
 NOBLOCK 
 

Figure 476. BLOCK Format

Tells whether or not to print the graphic portion of the HISTOGRM for block
length.

MFX for z/OS Programmer’s Guide 15–3

Control Parameters for HISTOGRM

REC

 REC 
 
 
 NOREC 
 

Figure 477. REC Format

Tells whether or not to print the graphic portion of the HISTOGRM for record
length.

BIGSTOP

 BIGSTOP 
 
 
 NOBIGSTP 
 

Figure 478. BIGSTOP Format

Tells whether or not to terminate the HISTOGRM run if an RDW value greater
than the DCB LRECL is encountered in the input file.

15–4 MFX for z/OS Programmer’s Guide

 Job Control Language

Job Control Language

The following example shows a sample execution of HISTOGRM.

1. SYSUT1 is the variable-length file to be scanned. Specify the DCB parameter if

//L6L7 JOB
//STEP1 EXEC PGM=HISTOGRM
//STEPLIB DD DSN=HISTOGRM,DISP=SHR
//SYSUT1 DD UNIT=3490,VOL=SER=000001, 1
// DSN=VLRECS,LABEL=(1,SL),
DISP=OLD
//SYSPRINT DD SYSOUT=A 2
//SYSIN DD * 3
KEYL=50,DEVWK=3390,NOBLOCK,NOBIGSTP
/*

Figure 479. Sample JCL/Control Stream for HISTOGRM

SYSUT1 is a non-standard label tape.

2. SYSPRINT is the data set on which printed output will appear. The DCB (not
illustrated) is: DCB=(LRECL=121,BLKSIZE=121,RECFM=F).

3. You may use DD DUMMY instead of SYSIN DD *. Specify //SYSIN DD

DUMMY,DCB=(LRECL=80,RECFM=FB,BLKSIZE=80).

Executing HISTOGRM through an E15 Exit

It is possible to execute HISTOGRM during a sort by specifying an E15 exit in the
MODS control statement and coding HISTE15 as the r value. This produces a
printout of the HISTOGRM for Records at the conclusion of the job. (It is, however,
not possible to get a printout of the HISTOGRM for Blocks when initiating
HISTOGRM in this way.)
The following example shows a sample execution of HISTOGRM by an E15 exit
during a sort.

MFX for z/OS Programmer’s Guide 15–5

Executing HISTOGRM through an E15 Exit

//HISTSORT JOB
//STEP2 EXEC PGM=SYNCSORT
//SORTIN DD UNIT=3490,VOL=SER=000001, 1
// DSN=VARDATA,LABEL=(1,SL),
// DISP=OLD
//SORTOUT DD UNIT=3490,VOL=SER=000002, 2
// DSN=SORTED.DATA,LABEL=(1,SL),
// DISP=(,KEEP)
//SORTWK01 DD UNIT=SYSDA,SPACE=(CYL,(20,10)) 3
//SORTWK02 DD UNIT=SYSDA,SPACE=(CYL,(20,10))
//SORTWK03 DD UNIT=SYSDA,SPACE=(CYL,(20,10))
//SYSOUT DD SYSOUT=A 4
//MODLIB DD DSN=SYS1.SYNCLIB,DISP=SHR 5
//SYSIN DD * 6
SORT FIELDS=(4,10,CH,A)
MODS E15=(HISTE15,7400,MODLIB,N) 7
//SYSPRINT DD SYSOUT=A 8
//HISTIN DD * 9
WIDTH=40
/*

Figure 480. Sample JCL/Control Stream for HISTOGRM Initiated by an E15 Exit

1. SORTIN is a DD statement for MFX. It contains the data set that will be analyzed
while it is being sorted. The data set name is VARDATA, and it is found on the standard
labeled tape with the volume serial number 000001. The data set is already in
existence. If SORTIN is not a standard label tape, DCB parameters must be specified.
Note that RECFM must be either V, VB, or VBS.

2. SORTOUT is a DD statement for MFX. It assigns the data set name SORTED.DATA to
the output file, and specifies a 3490 tape unit with the volume serial number 000002. It
is not yet in existence. The DCB parameters default to those of SORTIN.

3. SORTWK01, SORTWK02, and SORTWK03 are DD statements for MFX. They reserve
20 cylinders of primary space, 10 cylinders of secondary space on direct access devices
for intermediate storage.

4. SYSOUT is a DD statement for MFX. It assigns the MFX messages to the output device
associated with class A.

5. The MODLIB DD statement is used to define the partitioned data set in which the
HISTE15 program resides; MODLIB is referenced in the MODS control statement. The
data set name is SYS1.SYNCLIB, and the DISP shows the library may be shared.

6. The SYSIN DD * statement marks the beginning of the input stream that includes the
sort control statements. The SORT control statement shows that one control field will
be sorted on. It consists of bytes 4-13 of the record, contains character data, and is to be
sorted in ascending order.

15–6 MFX for z/OS Programmer’s Guide

 Executing HISTOGRM through an E15 Exit

The MODS control statement must specify an E15 exit as an exit-type parameter
and give HISTE15 as the exit routine name. HISTE15 takes 5000 bytes of storage
and resides in the main MFX library referenced here by a DD statement named
MODLIB. The routine does not require link-editing during sort execution.

7. SYSPRINT is the data set on which the printout from HISTE15 appears. Its DCB is:
DCB=(LRECL=121,BLKSIZE=121,RECFM=F).

8. The HISTIN DD statement is optional. It is used to override any default values. The
following DCB parameter must be specified: DCB=(LRECL=80,RECFM=
FB,BLKSIZE=80). (With HISTIN DD *, the DCB is not necessary.)

Defaults for HISTE15

NRECS= ALL
WIDTH= 20
DEVWK= The same SORTWK devices used in executing this sort.
KEYL= End of key field furthest into record for this sort.
BIGREC= 0 (This cannot be overridden.)
NOBLOCK (This cannot be overridden.)
REC
NOBIGSTP (This cannot be overridden.)

MFX for z/OS Programmer’s Guide 15–7

 1 HIS007I NUMBER OF BLOCKS................. 898

15–8
HIS008I TOTAL LENGTH OF ALL BLOCKS....... 104119
HIS009I AVERAGE BLOCK LENGTH............. 116
2 BLOCK
BLOCK LENGTH
COUNT MIN MAX
..........................................................................................
3 18 .****** . 40 59
80 .************************** . 60 79
145 .************************************************ . 80 99
205 .******************************************************************** . 100 119
253 .************************************************************************************ . 120 139
197 .***************************************************************** . 140 159
..........................................................................................
Executing HISTOGRM through an E15 Exit

yyyyyyyyyyy

cbddbdbdbdbdbdbdbffgfg

Sample Contents of HISTOGRM Output

1. HISTOGRM informational messages for blocks are printed at the top of the report. For explanations, see individual messages in the
message section which follows these examples.

2. BLOCK COUNT gives the number of blocks falling within the minimum and maximum numbers shown as BLOCK LENGTH. The
range is the WIDTH value that has been specified.

3. The asterisks are the graphic representation of the number of blocks within the range of block lengths.

MFX for z/OS Programmer’s Guide

 1 HISO10I NUMBER OF RECORDS.............. 952
HISO11I TOTAL LENGTH OF ALL RECORDS.... 93412
HIS012I AVERAGE RECORD LENGTH.......... 98
HIS016I KEY LENGTH..................... 50
HIS015I AVERAGE SPACE PER RECORD - L6.. 129
HIS014I RECOMMENDED SEG. SIZE - L7..... 72
HIS017I LINE WIDTH..................... 20
HIS018I LONGEST RECORD................. 155
HISO19I SHORTEST RECORD................ 40
HIS005I RECORDS TOO LONG............... 48
HIS023I RECORDS TOO SHORT.............. 41
HIS020I DEVICE TYPE.................... 3390
2 RECORD
RECORD LENGTH
COUNT MIN MAX

MFX for z/OS Programmer’s Guide

.........................................................................................
3 104 .********************************** . 40 59
181 .************************************************************ . 60 79
199 .***************************************************************** . 80 99
196 .**************************************************************** . 100 119
211 .********************************************************************* . 120 139
61 .******************** . 140 159
.........................................................................................

Sample Contents of HISTOGRM Output

1. HISTOGRM informational messages for records are printed at the top of the report. For explanations, see individual messages in
the message section which follows these examples.

2. RECORD COUNT gives the number of records falling within the minimum and maximum numbers shown as RECORD LENGTH.
The range is the WIDTH value that has been specified.

3. The asterisks are the graphic representation of the number of records within this range of record lengths.

15–9
Executing HISTOGRM through an E15 Exit
Executing HISTOGRM through an E15 Exit

HISTOGRM Messages
HISnnnA messages indicate a critical error condition. HISTOGRM ter-
minates to allow you to correct the error(s) so that a success-
ful program may be run.

HISnnnI messages are informational or indicate a non-critical error.

They are printed on the HISTOGRM output for blocks and
records and contain statistical information inserted by HIS-
TOGRM.

HIS001A INVALID CONTROL CARD

EXPLANATION: A blank control statement or an incomplete
control parameter was found.

HIS002A INVALID DATA ON CONTROL CARD

EXPLANATION: An invalid control parameter was found.

HIS003A EXPECTED CONTIN NOT FOUND

EXPLANATION: A control statement continuation was indi-
cated either by a non-blank character in column 72 or by a
comma immediately after the last control field, but no contin-
uation card image was found.

HIS004A INVALID DCB OR ACB DATA

EXPLANATION: For HISTOGRM: The SYSUT1 data set was
opened and one of three errors was detected: (1) LRECL was
not specified, (2) BLKSIZE was not specified, (3) RECFM was
not V, VB, or VBS, or (4) the data set is a VSAM RRDS. For
HISTE15: The SORTIN record format was not variable-
length.
ACTION: For HISTOGRM: Check for a missing DCB parame-
ter if SYSUT1 is a non-standard label tape. If the file is a
standard label tape or a disk file, one of the DCB subparame-
ters may be missing, or the file is not a variable-length file.
For HISTE15: Ensure that SORTIN is a variable-length data
set.

HIS005I RECORDS TOO LONG nnnn

EXPLANATION: Records with lengths exceeding the length
specified in the DCB were found. The nnnn represents the
number of long records found. Long records have no effect on
other HISTOGRM statistics.

15–10 MFX for z/OS Programmer’s Guide

 Executing HISTOGRM through an E15 Exit

HIS006A INVALID SPAN CONTROL FIELD BLOCK nnnn LOGI-

CAL RECORD nnnn {DATA SET # nnnn}
EXPLANATION: The third byte of the four byte record
descriptor word preceding a variable-length record does not
contain a valid code X'00', X'01', X'10', X'11', or the code is
inconsistent with the code of the previous segment. The block
and record number being processed are included in the mes-
sage text. The first 100 bytes of both current and previous
segment along with their RDWs, follows this message. DATA
SET # will be the concatenation number within SYSUT1 if
the input data set is concatenated.

HIS007I NUMBER OF BLOCKS nnnn

EXPLANATION: The total number of blocks read from the
SYSUT1 data set is given on the HISTOGRM for blocks.

HIS008I TOTAL LENGTH OF ALL BLOCKS nnnn

EXPLANATION: The total length in bytes of all blocks read is
given on the HISTOGRM for blocks.

HIS009I AVERAGE BLOCK LENGTH nnnn

EXPLANATION: The average block length of all blocks read
is given on the HISTOGRM for blocks.

HIS010I NUMBER OF RECORDS nnnn

EXPLANATION: The total number of records read from the
SYSUT1 data set is given on the HISTOGRM for records. The
total will exclude any records with lengths greater than the
length specified in the DCB.

HIS011I TOTAL LENGTH OF ALL RECORDS nnnn

EXPLANATION: The total length in bytes of all records read
is given on the HISTOGRM for records.

HIS012I AVERAGE RECORD LENGTH nnnn

EXPLANATION: The total length of all records is divided by
the number of records and the quotient is given on the HIS-
TOGRM for records.

HIS013I NUMBER OF SPANNED RECORDS

EXPLANATION: The number of records contained within two
or more blocks is given on the HISTOGRM for records.

MFX for z/OS Programmer’s Guide 15–11

Executing HISTOGRM through an E15 Exit

HIS014I RECOMMENDED SEG. SIZE - L7

EXPLANATION: The recommended segment size is given on
the HISTOGRM for records. Supply MFX with this value
either through L7 in the PARM field of the EXEC statement
or through l7 in the LENGTH parameter of the RECORD con-
trol statement.

Note: If the recommended number is 0, the range of record lengths

in the file was too wide to compute an optimal value. In this case, do
not supply an L7.

HIS015I AVERAGE SPACE PER RECORD - L6

EXPLANATION: The average work space necessary for each
record is given on the HISTOGRM for records. Supply MFX
with this value either through L6 in the PARM field of the
EXEC statement or through l6 in the LENGTH parameter of
the RECORD control statement.

Note: If the recommended number is 0, the range of record lengths

in the file was too wide to compute an optimal value. In this case, do
not supply an L6.

HIS016I KEY LENGTH nnnn

EXPLANATION: The end location of the last control field in
the record is given on the HISTOGRM for records.

HIS017I LINE WIDTH nnnn

EXPLANATION: The numeric interval between the mini-
mum and maximum block/record length is given on the HIS-
TOGRM for records.

HIS018I LONGEST RECORD nnnn

EXPLANATION: The length of the longest record read; that
is the record containing the largest value in the record
descriptor word.

HIS019A INVALID DEVICE TYPE

EXPLANATION: An invalid device type was specified on the
control statement in the SYSIN data set.

HIS019I SHORTEST RECORD nnnn

EXPLANATION: The length of the shortest record read is
given on the HISTOGRM for records.

15–12 MFX for z/OS Programmer’s Guide

 Executing HISTOGRM through an E15 Exit

HIS020I DEVICE TYPE nnnnnn

EXPLANATION: The type of intermediate storage device to
be used for the sort is given on the HISTOGRM for records.

HIS021I BLOCK PARAMETER IGNORED

EXPLANATION: Information about blocks cannot be col-
lected when running HISTE15 during a sort.

HIS022A INPUT FILE IS EMPTY

EXPLANATION: There are no records in the input file which
are not longer than the data set’s LRECL.

HIS023I RECORDS TOO SHORT nnnn

EXPLANATION: Records with lengths less than the KEYL
value specified for the HISTOGRM execution were found. The
nnnn is the number of short variable-length records in the
file.

HIS024I LRECL nnnnn,BLKSIZE nnnnn,RECFM xxx

EXPLANATION: The logical record length, block size, and
record format of the input data set obtained from the SYSUT1
DCB after OPEN.

HIS025A INVALID RDW/RECORD LENGTH BLOCK nnnn LOGI-

CAL RECORD nnnn {DATA SET # nnnn}
EXPLANATION: The RDW value of the current record is
greater than the DCB LRECL and HISTOGRM has been
requested to terminate (thru the BIGSTOP parameter). The
block and record number are supplied in the message and the
first 100 bytes of the record and the RDW follow the message.
DATA SET # will be the concatenation number within SYS-
UT1, if the input data set is concatenated.

HIS025I INVALID RDW/RECORD LENGTH BLOCK nnnn LOGI-

CAL RECORD nnnn {DATA SET # nnnn}
EXPLANATION: The RDW value of a record is greater than
the DCB LRECL. Block number and record number are sup-
plied in the message text, along with the concatenation num-
ber if SYSUT1 is concatenated.

MFX for z/OS Programmer’s Guide 15–13

Executing HISTOGRM through an E15 Exit

HIS026I INPUT DATA SET IS VSAM ... NO BLOCK STATISTICS

GATHERED
EXPLANATION: The input to HISTOGRM is a VSAM data
set; therefore block statistics are not produced for this HIS-
TOGRM execution.

HIS027A SYSUT1 DD STATEMENT MISSING

EXPLANATION: The input data set is absent; the HIS-
TOGRM run has terminated.

HIS028A VSAM LOGICAL ERROR nn

EXPLANATION: An error occurred while reading a VSAM
data set. For the definition of the error number, nn, consult
one of the following IBM publications:
• DFSMS/MVS Macro Instructions for Data Sets, SC26-
4913

HIS029A VSAM OPEN ERROR nn

EXPLANATION: An error occurred during an attempt to
OPEN a VSAM file. For the definition of the error number, nn,
consult one of the following IBM publications:
• DFSMS/MVS Macro Instructions for Data Sets, SC26-
4913

HIS030A message text

EXPLANATION: An I/O error has occurred. The message text
gives a detailed description of the error.

HIS031A INVALID BDW ENCOUNTERED BLOCK nnnn

EXPLANATION: The block descriptor word for block number
nnnn, was either zero or greater than the DCB blocksize.

15–14 MFX for z/OS Programmer’s Guide

Chapter 16 Value-Added Products

This chapter describes the Syncsort MFX value-added products:

• Syncsort PROCSort – An Accelerator for SAS® Sorting
• Syncsort PipeSort
• Syncsort ZPSaver
These products significantly improve performance and enhance programmer
productivity.

Syncsort PROCSort – An Accelerator for SAS® Sorting

Syncsort PROCSort is a high performance replacement for the SAS-provided
procedure PROC SORT. Compared to PROC SORT, Syncsort PROCSort reduces the
resources required for sorting within SAS applications and significantly cuts sort
elapsed time.
Sort processing within SAS often consumes as much as 30 percent of CPU time and
EXCPs. Because sorting is such a large part of system activity, Syncsort
PROCSort’s efficiency results in noticeable improvements in overall system
throughput. This reduced elapsed time from Syncsort PROCSort makes it possible
for SAS applications to complete much faster.
Syncsort PROCSort improves performance by providing a direct interface between
MFX and SAS. This frees MFX to use its high performance techniques –
sophisticated access methods, path length minimization algorithms and I/O
optimization.
No modifications to MFX are required to install and use Syncsort PROCSort.

Syncsort MFX Programmer’s Guide 16–1

 For more detailed information regarding the use and installation of Syncsort
PROCSort, refer to the booklet titled Syncsort PROCSort Installation and Use
Guide.

Syncsort PipeSort
Syncsort PipeSort works with MFX to run multiple sorts simultaneously on the
same input data. For large input files, Syncsort PipeSort significantly reduces total
elapsed time compared to running separate sort jobs.
Syncsort PipeSort reads SORTIN once and distributes the input records to up to
eight simultaneous MFX executions. The complete range of MFX control
statements and PARMs is available for the individual sort operations.
The output files are differently sequenced according to user-specified sort keys and
are written to different SRTnOUT DD data sets.
Optionally, you can use an inline E15 exit, with or without one or more E35 exits.
An inline E15 input exit can supply the input data to Syncsort PipeSort, and E35
output exits can accept the different output record sets.
For detailed information regarding installation and implementation through z/OS
JCL, refer to the Syncsort PipeSort User’s Guide.

Syncsort ZPSaver
Syncsort ZPSaver is a set of enhanced technologies for MFX to offload copy, SMS
compression and sort processing to zIIP processors, effectively reducing the
workload on the main CPU. Specifically, Syncsort ZPSaver includes:
For more detailed information regarding the use and installation of Syncsort
ZPSaver technologies, refer to the Syncsort ZPSaver User’s Guide.

Syncsort ZPCopy
Syncsort ZPCopy is used with SORT FIELDS=COPY
Benchmark tests show significant performance advantages when using Syncsort
ZPCopy:
• Significantly reduced TCB CPU time – up to 95%
• Significantly reduced elapsed time – up to 25%

16–2 Syncsort MFX Programmer’s Guide

Syncsort ZPCompress
Syncsort ZPCompress is used in most use cases when a sort or copy has SMS
version 1 compression attributes on SORTIN or SORTOUT (non-VSAM, non-
OUTFIL) data sets.
When reading or writing SMS-compressed data sets, CPU time typically increases
significantly while DASD usage and elapsed time are reduced dramatically.
Syncsort ZPCompress reduces the CPU time of compression/decompression by
offloading this processing to zIIP engines – performing 100% compatible
decompression on SORTIN and 100% compatible compression for SORTOUT while
achieving the equivalent reductions in DASD usage and elapsed time.
Benchmark tests for Syncsort ZPCompress demonstrate:
• Between 80% and 90% reduction in CPU time

Syncsort ZPSort
Syncsort ZPSort is initiated with SORT FIELDS=(x,y,ZZ,A…)
Benchmark tests show significant performance advantages when using Syncsort
ZPSort:
• Significantly reduced TCB CPU time – up to 95%

Syncsort ZPEncrypt
Syncsort ZPEncrypt is initiated with SWCRYPT =(YES)
Benchmark tests show significant performance advantages when using Syncsort
ZPEncrypt in the Syncsort ZPSort application:
• Significantly reduced TCB CPU time – up to 50%

Syncsort MFX Programmer’s Guide 16–3

16–4 Syncsort MFX Programmer’s Guide
Chapter 17 Messages

Syncsort MFX Messages

All messages issued by MFX have the form:
WERnnnx Message text
where nnn is the message number and x may be any of the letters A through I. The
interpretation of the suffix letter x is given below.
A (action) messages indicate a critical error condition: MFX terminates in order
to allow you to correct the error(s) so that a successful sort/merge may be run.

Example WER012A NO FLD DEFINER

B (tuning) messages provide information that may be useful in adjusting the job/
control stream to the actual demands of the job.

Example WER151B SECONDARY EXTENTS OBTAINED xxx

C-I (informational) messages document decisions internal to the sort as well as

MFX’s response to error conditions that are not severe enough to warrant sort/
merge termination.

Example WER177I TURNAROUND SORT PERFORMED

WER185I SORTIN DCBBLK GT ACTUAL, I/O INEFF

MFX provides an interactive message explanation facility, SS31MSG, that gives

online access to all MFX message texts and their explanations by message number.
If SS31MSG is included as an option in a PDF menu, you can invoke it from that
menu. Otherwise, you may invoke SS31MSG from the command line of any ISPF
panel by entering this command:

Syncsort MFX Programmer’s Guide 17–1

 TSO %SS31MSG

Figure 481. Command to Invoke SS31MSG

The installation of the SS31MSG facility is optional. Therefore, if you are unable to
invoke the facility as described, you should contact your system administrator for
more information.
Note: All messages that refer to SORTIN, SORTWK, SORTOF, and SORTOUT
provide the actual DD name, which reflects any changes made via a DD name over-
ride or a prefix override.

WER001A COL 1 OR 1-15 NOT BLANK

EXPLANATION: This message is triggered by a character in
column 1 of the END control statement or in columns 1-15 of a
continuation statement following a statement with a character in
column 72, or by a non-blank character in columns 1-15 of a sort
control statement in the $ORTPARM data set. These columns
must be left blank.

WER002A EXCESS CARDS

EXPLANATION: The static internal storage area is inadequate
for the quantity and/or complexity of the control statements in
this application. Either the minimum storage value set at
installation time is too low, or insufficient storage is available in
your region.
ACTION: Ask the systems programmer in charge of MFX
installation to increase the minimum storage (MINCORE) value
unless the storage available in the region is less than the
minimum storage value. In that case, increase the storage
available in the region or partition so that it at least equals the
minimum storage value.

WER012A NO FLD DEFINER

EXPLANATION: The FIELDS operand was not specified on the
SORT/MERGE control statement.

WER017A ERR IN DISP LENGTH VALUE

EXPLANATION: The length and displacement value of a control
field is greater than 4092 (4084 for variable-length records), or
less than one, or the sum of the lengths of all control fields
exceeds 4092 (4084 for variable-length records).

17–2 Syncsort MFX Programmer’s Guide

 WER018A CTL FLD ERR
EXPLANATION: An error was detected in the SORT/MERGE
control statement for the data format of a control field. The
format was specified for one field but not for another, or bit
comparisons were specified and FORMAT=BI was not specified.

WER026A L1 NOT GIVEN

EXPLANATION: The LENGTH operand on a RECORD control
statement does not contain an l1 value.

WER027A CONTROL FIELD BEYOND RECORD

EXPLANATION: The last byte of a SORT/MERGE or JOINKEYS
control field is located beyond the maximum record length
specified or column 32750, or a variable-length record is shorter
than the ending location of a specified SORT/MERGE or
JOINKEYS control field in an execution for which this is defined
as cause for MFX termination (see “VLTEST” on page 5-30).
Program HISTOGRM may be used to determine the length of the
shortest record in the input file.

WER029A IMPROPER EXIT

EXPLANATION: The set of legal exits depends on the sorting
technique chosen. A merge or copy may not specify any Phase 1
or Phase 2 exits; a copy may not specify exit E32 or E61; and a
sort or merge with data fields of Y2x or PD0 formats may not
specify exit E61.

WER032A EXIT E61 REQUIRED

EXPLANATION: A SORT control statement specified "E" in the
FIELDS parameter but program exit E61 was not specified on a
MODS control statement.

WER033A CONTROL FIELD COLLATING ORDER E REQUIRED

EXPLANATION: Program exit E61 was specified on the MODS
control statement but "E" was not specified in the FIELDS
parameter of the SORT control statement.

WER036B G=ggg, B=bbb, SEGLEN=sss, BIAS=zz

EXPLANATION: The tuning information displayed is as follows:

G=ggg ggg is the number of records that can be contained

in MFX’s working virtual storage area. For vari-
able-length records, this number is the number of
segments.

Syncsort MFX Programmer’s Guide 17–3

 B=bbb bbb indicates the physical blocking used for inter-
mediate storage. For fixed-length records, this
number represents the blocking factor. For vari-
able-length records, it represents the blocksize.
The B value will not appear in the message for
incore or turnaround sorts.

SEGLEN=sss This value appears in the message for variable-

length records, when the execution is not an incore
or turnaround sort. It reflects the segment length
used in MFX’s working storage during Phase 1.

BIAS=zz zz reflects the degree of prior sequencing in the

input data. The number displayed ranges from 00
to 99 indicating random to highly sequenced input.
The BIAS value is not included in the message for
an incore or turnaround sort, where it is 100 by
definition.

WER037A REXX ENVIRONMENT UNAVAILABLE

EXPLANATION: One or more REXX exits were specified in a
MODS control statement, but the required operating system and/
or TSO environment is not available.

WER039A INSUFFICIENT VIRTUAL STORAGE

EXPLANATION: The amount of virtual storage available to MFX
is not large enough to permit execution.
ACTION: Verify that virtual storage is specified properly. Check
that the region size is sufficient for execution.

WER044A EXIT Exx INVALID OPTION

EXPLANATION: The exit routine shown in the message
specified an invalid option for the modification of a DCB
parameter of a sort/merge data set.

WER046A SORT CAPACITY EXCEEDED

EXPLANATION: All available intermediate storage is
exhausted, including any secondary allocation. Sort processing
cannot continue.
ACTION: Supply more intermediate storage (see “Disk Sort
Intermediate Storage Calculation Formulas” on page 14-5) or use
the MAXSORT technique.

17–4 Syncsort MFX Programmer’s Guide

 WER047A RCD CNT OFF, IN x, OUT y
EXPLANATION: The actual number of records specified in the
SIZE parameter on the SORT control statement (the IN value)
was not equal to the number of records read from the input data
set (the OUT value). (This comparison is made only when the
SIZE parameter specifies an actual number of records.)
The actual number of records (the IN value) for FILSZ=n speci-
fied either on the SORT control statement or as a PARM option
was not equal to the total number of records (the OUT value)
from the input data set after any changes due to the INCLUDE/
OMIT control statement, and an E14 or E15 exit routine, and
SKIPREC, STOPAFT, or join processing.

WER048I E16 EXIT CALLED

EXPLANATION: Program exit E16 was entered after all
available SORTWORK space was exhausted.

WER049A SUM FIELD OVERFLOW

EXPLANATION: Summing of two equally keyed records could
not be done due to a numeric overflow or underflow in a defined
SUM field. The WER049A critical message is issued instead of
the WER049I warning message if the OVFLO=RC16 PARM or
the installation parameter SUMOVFL=RC16 is in effect.
ACTION: If complete summing is desired, use the INREC control
statement if possible to pad the fields with leading zeros of the
proper numeric format. Adjust the SUM field and other control
statement fields accordingly.

WER049I SUM FIELD OVERFLOW

EXPLANATION: Summing of two equally keyed records could
not be done due to a numeric overflow or underflow in a defined
SUM field.
ACTION: If complete summing is desired, use the INREC control
statement if possible to pad the fields with leading zeros of the
proper numeric format. Adjust the SUM field and other control
statement fields accordingly.

WER050I SUM CONTROL STATEMENT IGNORED

EXPLANATION: A SUM control statement was specified in a
SORT FIELDS=COPY application. Since a COPY operation does
not use SORT/MERGE key fields, the specification of SUM,
which operates on equally keyed records, is illogical.
ACTION: The SUM statement will be ignored, but the
application should be checked for correct specification of control
statements.

Syncsort MFX Programmer’s Guide 17–5

 WER052I END SYNCSORT - jobname, stepname, procstepname,
DIAG=hhhh,hhhh,...
EXPLANATION: MFX has successfully completed execution. The
hexadecimal information following the DIAG keyword is likely to
change from execution to execution. It is internal diagnostic
information intended for use by Syncsort MFX personnel in
Precisely Support.

WER054I RCD IN x, OUT y

EXPLANATION: For a non-join application, the x represents the
number of records read from the input data set(s). For a join
application, the x represents the resulting number of records
created by the join processing. If OUTFIL statements are not
present, the y represents the number of records in the output file.
If OUTFIL statements are present, the y represents the number
of records available for OUTFIL processing.

WER055I INSERT x, DELETE y

EXPLANATION: The x represents the number of records
inserted by user exit routines. The y represents the number of
records deleted by user exit routines, SUM, and INCLUDE/OMIT
control statements.

Note: For MAXSORT, these counts are cumulative for the entire MAX-
SORT application.

WER059A RCD LNG INVALID FOR DEVICE

EXPLANATION: The logical record length specified for a fixed-
length input data set plus overhead, if any, is too large to fit on
one disk track of the intermediate storage device

WER061A I/O ERR jobname, stepname, unit address, device type,

DDname, operation attempted, error description, last
seek address or block count, access method.
EXPLANATION: An I/O error has occurred on the device whose
address is given. I/O errors are often transient - resubmitting the
job may result in a successful run. However, if the I/O error is on
an input DD, the following should be checked first:

• When the input consists of concatenated data sets, check that the
largest blocksize is available at sort initialization. See
“Concatenating Input Data Sets” on page 4-6.

• If the data set is on disk and has just been created by another
program, check that this program opened the data set even if no
data was written to the file. The data set must be opened in order for
an end-of-file mark to be written. (In the absence of an end-of-file

17–6 Syncsort MFX Programmer’s Guide

 mark, MFX will tend to read whatever was on the disk as part of the
input data set, causing an I/O error.)

WER063A xxxxxx OPEN ERR

EXPLANATION: The data set shown cannot be successfully
opened.
ACTION: Check for missing DD statements.

WER065A DECK STRUCTURE ERROR

EXPLANATION: The end of the SYSIN data set was reached
before all user exit routines were read or an object deck was
missing its first statement.

WER066A APPROX RCD CNT x

EXPLANATION: Sort capacity was exceeded, so the sort
terminated. The approximate number of records processed by
MFX up to this point is given.

WER068A OUT OF SEQ SORTINxx[, BLOCK y]

EXPLANATION: A record in the SORTIN data set indicated by
xx is out of sequence according to the FIELDS specification on the
MERGE statement. The number y of the block containing the
out-of-sequence record is given if an E32 exit was not used.

WER069A E39/OUTFIL INCOMPATIBLE

EXPLANATION: The E39 exit facility may not be utilized in
sorts/merges which also specify OUTFIL control statements.

WER070A ddname {TOTAL,SUBTOTAL,AVG,SUBAVG} FIELD

OVERFLOW
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. An overflow condition was generated during the
TOTAL, SUBTOTAL, AVG, or SUBAVG OUTFIL function for the
specified output file. All fields are totaled internally as 16-byte
PD values, allowing for 31 decimal digits. A field that is to be
converted to BI or FI format may not have a value greater than
20 decimal digits.
ACTION: This error is usually due to invalid data in the specified
fields. Check the fields specified in the indicated parameter and
check the actual data in those fields to ensure that no total will
exceed 31 decimal digits. In some cases, within TRAILER2 or
TRAILER3, you could change SUBTOTAL to TOTAL or SUBAVG
to AVG to reduce the possibility of overflow.

Syncsort MFX Programmer’s Guide 17–7

 WER071A MAXIMUM NUMBER OF RECORDS EXCEEDED
EXPLANATION: MFX’s default internal limit on the maximum
number of records that can be sorted has been exceeded. By
default, the internal limit on the number of records that can be
processed for variable-length data or for a sort application that
uses the EQUALS option is 4,294,967,295 records. Specify the
EXTCOUNT PARM to increase the internal limit to
140,737,488,355,327 records. Fixed-length sorts without
EQUALS have automatic support for the maximum number of
records allowed by the EXTCOUNT PARM. For additional
information, see “EXTCOUNT” on page 5-14.

WER072I {EQUALS, NOEQUALS} [,RESET] {,BALANCE, ELAP, CPU,

IO} IN EFFECT
EXPLANATION: This message indicates the status of three MFX
options that could affect how your data was processed. The status
of the EQUALS option, primarily used to retain the sequence of
equally-keyed input records, is given first. “RESET” is indicated
if the RESET option was used to prevent VSAM from treating
output data sets created with the REUSE option as MOD data
sets. The last parameter is the optimization mode in effect, i.e.,
BALANCE, ELAP, CPU, or IO.

WER073I ddname: dsname [(FIRST of n)]

EXPLANATION: The ddname will be SORTIN, SORTJNF1,
SORTJNF2 or SORTMInn. This informational message displays
the input DD data set name. For concatenated DDs, only the first
data set name will be displayed and the number of
concatenations will be displayed in “FIRST of n”. For MULTIIN
input, WER073I will be displayed for only the first input ddname
that is read.

WER074I ddname: DSNAME=dsname

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. This informational message displays the output DD
data set name. This message will be provided for each output file
specified.

WER101D INVALID TAPE TYPE IN PARM FIELD

EXPLANATION: An invalid tape type was specified for DEVIN/
DEVOUT in the PARM field of the EXEC statement. MFX
ignored the invalid parameter.

17–8 Syncsort MFX Programmer’s Guide

 WER102A COBEXIT=COB2 AND COBOL E15 AND E35 EXITS
FOUND IN COPY APPLICATION
EXPLANATION: A COBOL E15 and COBOL E35 may not both
be specified in a copy application if the COBEXIT=COB2
installation option is in effect. Only one of the exits is permitted.

WER103D INVALID MESSAGE TYPE IN PARM FIELD

EXPLANATION: An invalid message code was specified in the
PARM field of the EXEC statement or in the invoking program
parameter list. MFX ignored the invalid parameter.

WER104A REXX E15 AND REXX E35 EXITS FOUND IN A COPY

APPLICATION
EXPLANATION: A REXX E15 and a REXX E35 may not both be
specified in a copy application. Only one of the exits is permitted.

WER105A INCOMPATIBLE LEVELS BETWEEN THE STATIC AND

DYNAMIC LIBRARIES OF A COBOL OR C EXIT
EXPLANATION: When using either a C or COBOL exit, insure
that the run-time dynamic Language Environment libraries are
at the same or higher level than the libraries used for the compile
or link-edit of the exit.

WER106A ddname INVALID DEVICE TYPE

EXPLANATION: The ddname is SORTIN, SORTINnn,
SORTJNF1, SORTJNF2, SORTMInn, SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. This file resides on an invalid device type. Valid
device types include the IBM 3380, 3390, and 9345 direct access
devices and their equivalents as well as the IBM 3420, 3480,
3490, and 3590 series tape devices and their equivalents.

WER107A ddname RECFM INCOMPATIBLE WITH REPORT

WRITING
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The RECFM specified for the file did not include the
'A' (ASA control character) specification that is required when
report writing is requested.

WER108I ddname: RECFM= ;LRECL= {;BLKSIZE=,CISIZE=} [;CINV

ACCESS]
EXPLANATION: The ddname will be SORTIN, SORTMInn,
SORTJNF1, or SORTJNF2. This informational message lists the
DCB characteristics used by MFX to process the input file. For a
non-VSAM data set that is concatenated, the DCB characteristics

Syncsort MFX Programmer’s Guide 17–9

 are for the first of the concatenated data sets, except for
BLKSIZE, which is the largest of all data sets in the
concatenation examined at sort initialization time. For a VSAM
data set, the CISIZE is provided; if control interval access was
used, the CINV ACCESS portion of the message will be
displayed.

WER109I MERGE INPUT: TYPE={F,V};LRECL=

EXPLANATION: This informational message lists the DCB
characteristics used by MFX to process the input files for a
merge.

WER110I ddname RECFM= ;LRECL= {;BLKSIZE=,CISIZE=} [;CINV

ACCESS]
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. This informational message lists the DCB
characteristics used by MFX to process the indicated output file.
This message will be provided for each output file specified. For a
VSAM data set, the CISIZE is provided; if control interval access
was used, the CINV ACCESS portion of the message will be
displayed.

WER111A [ddname] {INREC,OUTREC,TOTAL/SUBTOTAL,MIN/

SUBMIN, MAX/SUBMAX,AVG/
SUBAVG} INVALID DATA CONVERSION REQUESTED
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. Data conversion has been requested for INREC,
OUTREC, OUTFIL OUTREC, TOTAL/SUBTOTAL, etc., as
indicated, and one of the following error conditions has occurred:
1. The length of the field to be converted is too large.
2. Data conversion has been requested for a field that is not
specified as BI, CSF/FS, FI, PD, Y2ID, Y2IP or ZD.
3. Invalid or conflicting EDIT/SIGNS parameters were
specified.

WER112A INVALID VALUES IN FIELD PARAMETER

EXPLANATION: An invalid value was specified in the FIELDS
operand of the SORT/MERGE control statement.

WER113A TOO MANY SORT FIELDS

EXPLANATION: The number of sort control fields specified
exceeds the internal limits of the product. The absolute upper
limit on the number of sort control fields is 128; however,
depending on the complexity of an application, the limit may be

17–10 Syncsort MFX Programmer’s Guide

 reduced. When locale processing is used, the number of allowable
CH control fields is also limited by the length of those fields.

WER115A INVALID MOD NAME

EXPLANATION: An invalid name for a program exit was entered
on a MODS control statement.

WER116A THE FOLLOWING H/T IS GT LRECL:

EXPLANATION: This message flags any HEADERs, TRAILERs
or IFTRAIL TRLUPD records that exceed the LRECL
specification. The HEADER or TRAILER in error will be printed
on the next line. To correct this problem, see “Rules for Specifying
HEADER Subparameters” on page 2-104 or “Rules for Specifying
TRAILER Subparameters” on page 2-117.

WER117A INVALID ANSI CONTROL CHARACTER FOUND

EXPLANATION: An invalid ANSI control character appears in a
HEADER or TRAILER. The ANSI Control Character Table lists
the valid characters accepted by MFX.

WER117I INVALID ANSI CONTROL CHARACTER FOUND

EXPLANATION: An invalid ANSI control character appears in
an output data record. The sort will process the record as if a
blank control character had been found. This message will be
issued only once regardless of how many data records have
invalid ANSI characters. The ANSI Control Character Table lists
the valid characters accepted by MFX.

WER118A ddname INVALID OVERLAPPING FIELDS

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. An OUTFIL control statement contains a HEADER,
TRAILER or IFTRAIL TRLUPD parameter which contains
overlapping fields. This may be caused, for example, by a
positional subparameter specification which overlaps a
previously defined field.

WER119A NO DD NAME IN MODS FIELD

EXPLANATION: A DD name is missing on the MODS control
statement.

WER120A SEP. LKED NOT ALLOWED

EXPLANATION: A module for which separate link-editing was
specified on a MODS control statement is not allowed to be link-
edited separately.

Syncsort MFX Programmer’s Guide 17–11

 WER121A TASK CALL PARAM ERROR
EXPLANATION: If a 24-bit list is being used, either a control
statement address is zero, or the length of a control statement is
not positive, or the parameter list ends with the first word of a
two-word parameter. If a 31-bit list is being used, the last
parameter word in the list is not followed by the four byte field
X'FFFFFFFF'.

WER122A INVALID INTERMEDIATE STORAGE DEVICE

EXPLANATION: An invalid device was assigned as intermediate
storage. Valid devices include IBM’s 3380, 3390, and 9345 mass
storage system, and equivalent units.

WER123A IMPROPER RETURN CODE FROM Exx

EXPLANATION: An invalid return code was passed by the exit
that appears in the message. Valid return codes are 0, 4, 8, 12, 16
(and 20, if the exit is a COBOL or C E15 or E35).

WER124I [ESTIMATED] PREALLOCATED/USED SORTWORK

SPACE USAGE FACTOR {=,<,>}nn.nn
EXPLANATION: nn.nn represents the quotient obtained by
dividing the number of tracks assigned within preallocated
sortworks (sortworks allocated in the JCL or dynamically
allocated by an invoking program) by the number of tracks
actually used by MFX. The word ESTIMATED is included when
MFX’s derivation of this factor is inexact, for example, when all
sortwork data sets are not opened, or when data space or
hiperspace is used to contain part or all of the sortwork data.
Note that for MAXSORTs, the factor displayed is at or near "1.00"
for all but the last sort. For the last sort, the factor may be any-
where between "0.01" and "1.00" depending on the amount of
data sorted.

WER126I OUTFIL CONVERT WITHOUT THE OUTREC OR BUILD

PARAMETER MAY PRODUCE UNDESIRABLE OUTPUT
EXPLANATION: OUTFIL CONVERT requires the OUTREC or
BUILD parameter to produce consistent and blank-padded, if
appropriate, output records. If the OUTREC or BUILD
parameter is not specified with OUTFIL CONVERT, the sort may
still run to completion, but the output record will include the
input RDW in the first four bytes, and the record may be padded
to the output record length with residual storage contents that
are unpredictable in nature.

17–12 Syncsort MFX Programmer’s Guide

 WER130A I/O ERROR ON SYSIN
EXPLANATION: An I/O error occurred on SYSIN or
$ORTPARM.

WER131I PARM FIELD ERROR - xxxxxxxx

EXPLANATION: An invalid PARM was found in the PARM field
string that was passed to MFX. MFX terminated the PARM
processing by ignoring the remainder of this PARM string
without terminating the MFX application. The invalid PARM is
displayed in the message text if the PARM was passed on the
EXEC statement. If the invalid PARM was passed through the
$ORTPARM DD statement, the entire PARM string is written to
the SYSOUT data set, and an asterisk is displayed beneath the
invalid PARM.

WER133A Exx USER EXIT REQUESTED TERMINATION

EXPLANATION: Return code 16 was passed by the exit routine
shown in the message. MFX terminated.

WER135A TASK CALL/E35 TERMINATED PREMATURELY

EXPLANATION: An E35 exit routine (COBOL Output
Procedure) passed a return code of 8, terminating the sort before
the sort was able to pass all of the records. A SORTOUT data set
was not present.

WER135I TASK CALL/E35 TERMINATED PREMATURELY

EXPLANATION: An E35 exit routine (COBOL Output
Procedure) passed a return code of 8, terminating the sort before
the sort was able to pass all of the records. A SORTOUT data set
was not present.
This message may not indicate an error condition - it depends on
what the programmer intended. For example, this message will
be generated if a COBOL program using the SORT verb
RELEASEs 100 records in the Input Procedure without
RETURNing all 100 records in the Output Procedure because the
logic dropped to the bottom of the Output Procedure “prema-
turely.” If this is what the programmer intended, then no data
has been lost. If, however, the programmer intended the Output
Procedure to write all the records read in the Input Procedure,
then this message indicates a logic bug in the COBOL program.

WER136A {INREC,OUTREC,ddname OUTREC} HAS OVERLAPPING

FIELDS SPECIFIED
EXPLANATION: The column specification of a c: subparameter
in the indicated control statement overlaps a field previously
defined in the same control statement. Note that the

Syncsort MFX Programmer’s Guide 17–13

 subparameters used to define each field must be coded in the
order in which the fields will appear in the reformatted record.
The ddname will be SORTOUT, SORTOFxx, SORTOFx or the
ddname provided by an OUTFIL FNAMES parameter.

WER138A ddname BLKSIZE NOT EVENLY DIVISIBLE BY LRECL

EXPLANATION: The ddname is SORTIN, SORTINnn,
SORTJNF1, SORTJNF2, SORTMInn, SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. A block was read from the indicated file whose length
was not a multiple of the LRECL value, or the JCL or data set
attributes are incorrect.

WER141A ddname RECFM IS U

EXPLANATION: The ddname is SORTIN, SORTINnn,
SORTJNF1, SORTJNF2, SORTMInn, SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. MFX does not support undefined record format for
any of these files.

WER142A MIXED FIXED AND VARIABLE SORTxx TYPES NOT

SUPPORTED
EXPLANATION: MFX permits only one record format type (fixed
or variable) for input files per sort/merge. xx could be ‘IN’ or
‘MI’.

WER143A SORTIN LRECLS ARE MIXED

EXPLANATION: The LRECL must be the same for all fixed-
length files supplied to a merge.

WER144B UNEXPECTED VIRTUAL STORAGE FRAGMENTATION

EXPLANATION: The amount of virtual storage calculated by
MFX for Phases 2 or 3 was not available in a contiguous block.
Additional virtual storage was obtained to satisfy the sort
requirement. This condition was probably caused by virtual
storage not released by the user program in the job step (for
example, user exit buffer space was not released).

WER146B nnn BYTES OF EMERGENCY SPACE

EXPLANATION: The indicated amount of virtual storage has
been set aside by MFX for use by other programs (e.g., program
invoking the sort, system SVCs, tape management system).

17–14 Syncsort MFX Programmer’s Guide

 WER147I CONTROL FIELD GREATER THAN RECORD LENGTH,
POSSIBLE OUT OF SEQUENCE RECORD
EXPLANATION: The sort encountered a variable-length record
that was too short to contain all of the control field(s) specified in
the SORT statement. VLTEST instructed MFX to pad the record
with binary zeros to the length of the sort key and continue
processing. The added binary zeros account for the position of
this record in the sorted file, which may appear to be out of
sequence for this reason. The binary zeros are removed when the
record is processed for output. Program HISTOGRM may be used
to determine the length of the shortest record in the input file.

WER148A OPEN ERR SYSIN

EXPLANATION: SYSIN is either not present or cannot be
opened.

WER149B FRAGMENTED VIRTUAL STORAGE IN SORT PHASE

EXPLANATION: The virtual storage specified for MFX’s use was
not available in a contiguous block for Phase 1. This condition
was probably caused by a calling program or user exit routine.
MFX obtained its virtual storage in fragments and continued
execution. Note that the calling program or user exit routine used
virtual storage in such a way as to cause fragmentation, which
might another time result in ABEND 80A or S804.

WER151B SECONDARY EXTENTS OBTAINED xxx

EXPLANATION: This gives the number of secondary extents
obtained for SORTWKxx data sets.

WER152B REQUESTED VIRTUAL STORAGE NOT AVAILABLE, nnn

BYTES USED
EXPLANATION: The CORE parameter specified a value which
was not available when MFX received control. The number of
available bytes used by MFX is given.

WER153A INSUFFICIENT VIRTUAL STORAGE IN {INT.,FINAL}

MERGE PHASE
EXPLANATION: The amount of virtual storage available for the
indicated merge phase (the intermediate or final merge phase)
was not sufficient to allow execution.

WER154A NO MODS DD CARD

EXPLANATION: The DD statement whose name was specified
on the MODS control statement was not provided, so the user
exit routine cannot be found.

Syncsort MFX Programmer’s Guide 17–15

 WER157A SPANNED RECORD LENGTH LARGER THAN LRECL
EXPLANATION: A record from a VBS input data set contains a
record longer than the maximum record length specified by
LRECL in the DCB.
ACTION: Execute program HISTOGRM to get the length of the
longest record in the data set. Use this length for the LRECL
value in the DCB parameter of the input data set.

WER158I RECORD LENGTH GREATER THAN L2, CUT TO L2

EXPLANATION: A variable-length input record is longer than
the maximum record length specified by either LRECL in the
DCB or the l2 value in the RECORD control statement. (If l2 was
not specified, the variable-length record is longer than the l1
value.) MFX has truncated the record.
ACTION: For applications where MFX is reading SORTIN, if
truncation is not desired, execute program HISTOGRM to get the
length of the longest record in the data set. Use this length for
the LRECL value in the DCB parameter of the SORTIN data set.

WER159A RECORD LENGTH 0, SORTIN RECORD

xxx,xxx,xxx,xxx,xxx
EXPLANATION: An invalid variable-length record (length code
<4 in its Record Descriptor Word) has been found. If the record
was found in the input file, the number of the invalid record is
given. If the record was inserted from a user exit routine, the
number of the inserted record is given (for example, 45 indicates
the forty-fifth record read from the input file or inserted by a user
exit.)

WER160A RECORD LENGTH GREATER THAN LRECL,

TERMINATION REQUESTED
EXPLANATION: VLTEST has requested the sort to abort
because of the following condition. A variable-length record read
from the input file is longer than the maximum record length
specified by LRECL in the DCB or (after E15 processing) is
longer than the l2 value in the RECORD control statement. (If l2
was not specified, the l1 value was used as its default.)
ACTION: Change the LRECL or l2 value to reflect the record
length, or specify another value for VLTEST. Program
HISTOGRM may be used to determine the length of the longest
record in the input file.

WER161B ALTERNATE PARM USED

EXPLANATION: The alternate PARM option ($ORTPARM DD,
PARMEXIT or PARMTABLE) was used and MFX received the
parameters specified.

17–16 Syncsort MFX Programmer’s Guide

 WER162B ppp PREALLOCATED SORTWORK TRACKS, ddd
DYNAMICALLY ALLOCATED sss ACQUIRED IN xxx
SECONDARY EXTENTS, rrr RELEASED, TOTAL OF uuu
TRACKS USED
EXPLANATION: ppp is the number of tracks found available in
sortwork data sets which were allocated prior to MFX’s gaining
control. (These may have been allocated in the JCL or
dynamically allocated by an invoking program.) ddd is the
number of tracks dynamically allocated as primary space by
MFX. sss is the number of tracks acquired as secondary space, on
both preallocated data sets and data sets dynamically allocated
by MFX. xxx is the total number of secondary extents acquired.
rrr is the total number of unneeded tracks released from both
preallocated data sets and data sets dynamically allocated by
MFX. uuu is the total number of tracks actually used in sorting.
The following notes apply to the information in this message:

• ppp may not represent all of the preallocated tracks available, since
not all preallocated sortwork data sets may be opened by MFX.

• uuu may be less than the sum of ppp, ddd and sss since it
represents the space actually used and not the space available.

• For MAXSORTs, all dynamic allocation and secondary space

acquisition is done during the first sort. For this reason, the
WER162B message for the first sort will indicate the number of
tracks dynamically allocated, the number acquired via secondary
extents, etc. However, the WER162B message in all subsequent
MAXSORT sorts will report these tracks as “preallocated”.

WER164B www BYTES OF VIRTUAL STORAGE AVAILABLE, xxx

BYTES REQUESTED, yyy BYTES RESERVE
REQUESTED, zzz BYTES USED
EXPLANATION: The amount of virtual storage available (free)
when MFX received control is represented by w’s. The amount of
virtual storage requested for MFX’s use is represented by x’s. The
amount of virtual storage that the user requested MFX to reserve
below the 16-megabyte line is represented by y’s. The amount of
virtual storage used by MFX is represented by z’s. This message
reflects the total amount of virtual storage below and above the
16-megabyte line that was available to MFX and used by MFX.

WER165I STAT DATA REC NOT WRITTEN

EXPLANATION: The installation default for the MFX SMF
record feature is applied, but the sort did not invoke the module
that creates the sort statistical record. A possible reason for the

Syncsort MFX Programmer’s Guide 17–17

 sort’s not invoking the module may be that FREE=CLOSE was
coded on the SORTOUT (SYSUT2) or SORTWKxx DD statement.
ACTION: Remove the FREE=CLOSE parameter if full SMF
statistics are desired.

WER166I RECORD LENGTH GREATER THAN L3, CUT TO L3

EXPLANATION: MFX has truncated a variable-length record
prior to output processing. If an E35 exit was in use, the
truncated record was longer than the LRECL of the output file’s
DCB or greater than the l3 value on the RECORD control
statement. If an E35 exit was not in use, the truncated record
was longer than the LRECL in the output file’s DCB. If OUTFIL
processing is requested additional truncation may occur as a
result of the OUTFIL processing regardless of the action
requested by the VLTEST PARM.

WER167A RECORD LENGTH GREATER THAN OUTPUT LRECL,

TERMINATION REQUESTED
EXPLANATION: Prior to output processing MFX has
encountered a variable-length record longer than the l3 value on
the RECORD control statement (if an E35 exit was in use) or
longer than the LRECL in the output file’s DCB. Either the
VLTEST PARM requested MFX to terminate or OUTFIL
VLTRAIL was used.
ACTION: Change the LRECL in the output file’s DCB or the l3
value on the RECORD control statement (if an E35 exit is used)
to reflect the correct record length, or specify another value for
the VLTEST PARM.

WER168A CONTROL FIELD WITHIN RDW

EXPLANATION: A SORT/MERGE control field for a variable-
length file fell within the Record Descriptor Word of each record.
This is a critical error whenever the control field is specified with
a ZD or PD format code.

WER168I CONTROL FIELD WITHIN RDW

EXPLANATION: A SORT/MERGE control field for a variable-
length file falls within the Record Descriptor Word of each record.
(The first byte of the data portion of a variable-length record is at
byte position 5.)

WER169I RELEASE r.r BATCH nnnn TPF LEVEL n.n

EXPLANATION: Details on the release level, the batch number
from the base installation tape, and the last TPF applied to MFX
are given.

17–18 Syncsort MFX Programmer’s Guide

 WER170A CONCATENATED DATA SET, BLKSIZE NOT DIVISIBLE
BY LRECL
EXPLANATION: One of the files concatenated to a fixed-length
input data set or one of the multiple input files has a BLKSIZE
that is not evenly divisible by the original LRECL. Note that
‘CONCATENATED DATA SET’ could be replaced with
‘MULTIPLE INPUT’ in the message.
ACTION: Check the BLKSIZE specified on each of the DD
statements concatenated to the first DD statement of the input or
SORTMInn DD statements.

WER171A CONCATENATED DATA SET, LRECLS NOT EQUAL OR

RECFMS DIFFERENT
EXPLANATION: One of the files concatenated to a fixed-length
input data set or one of the multiple input files has an LRECL
not equal to the original LRECL; or one of the files concatenated
to a variable-length data set has an LRECL greater than the
original LRECL; or one of the files concatenated to a fixed or
variable-length data set has a RECFM not equal to the original
RECFM. Note that ‘CONCATENATED DATA SET’ could be
replaced with ‘MULTIPLE INPUT’ in the message.

WER172A CONCATENATED DATA SET, BLKSIZE GREATER THAN

ORIGINAL BLKSIZE
EXPLANATION: One of the files concatenated to an input data
set has a BLKSIZE greater than the original BLKSIZE.

WER173A BDW INVALID

EXPLANATION: The Block Descriptor Word of a block in the
input data set contains a value less than 8; or the Block
Descriptor Word contains a value greater than the number of
bytes actually read.
ACTION: Check the data set for the invalid block.

WER174A RDW INVALID, OVERFLOWS BUFFER

EXPLANATION: The Record Descriptor Word of a record in the
input data set is too large. (According to the RDW, the record
extends beyond the buffer.)
ACTION: Execute HISTOGRM to check the data set for an
invalid record.

WER175A INCORE SORT CAPACITY EXCEEDED

EXPLANATION: There are too many input records to fit in
virtual storage.

Syncsort MFX Programmer’s Guide 17–19

 ACTION: Either increase the amount of virtual storage the sort
is able to use or supply SORTWKxx DD statements. (The
DYNALLOC option may be used instead of SORTWKxx DD
statements.)

WER176A USER EXIT LKED FAILED

EXPLANATION: Exit routine(s) needing to be link-edited were
present, but the linkage editor passed a return code greater than
0.
ACTION: Check that the DD statement specified on the MODS
control statement is present in the JCL and contains the modules
specified on the MODS statement. Check that all exits (except
E11, E21, and E31) to be link-edited together have an external
name identical to the exit name. Examine the linkage editor
output for other errors.

WER177I TURNAROUND SORT PERFORMED

EXPLANATION: MFX was able to sort the input file without
using intermediate storage (SORTWKxx’s). All input data was
contained in virtual storage.

WER178A ddname [nnnnn] MEMBER NOT FOUND

EXPLANATION: An input DD statement specified a member of a
partitioned data set that could not be found. If a value nnnnn is
provided, it represents the concatenation number of the data set
that has the member-not-found condition.
ACTION: Check the DD statement for an error or list the
members of the partitioned data set.

WER179A ddname INVALID DCB PARAMETERS

EXPLANATION: The ddname is SORTIN, SORTINnn,
SORTJNF1, SORTJNF2, SORTMInn, SORTOUT, SORTOFxx,
SORTOFx, or the ddname provided by an OUTFIL FNAMES
parameter. MFX is unable to derive RECFM, LRECL, and
BLKSIZE parameters from the JCL, the DSCB on the disk or the
tape label.
ACTION: Check the JCL and the disk or tape labels for the error.

WER180A ddname MEMBER NOT SPECIFIED

EXPLANATION: The indicated input or output DD statement
defines a partitioned data set, but a member name has not been
specified.
ACTION: Specify a member name on the indicated DD statement
or change the partitioned data set to a sequential data set.

17–20 Syncsort MFX Programmer’s Guide

 WER182A INVALID RDW ddname BLOCK x
EXPLANATION: An invalid spanned record indicator was
detected in an input file whose RECFM=VBS, or an invalid
record length was detected in a copy operation. The block number
of the file is given.
ACTION: Execute HISTOGRM to check the data set for a record
containing invalid span bits. You can also use the VLTEST option
to turn off segment sequence checking if so desired.

WER183A SORTWORK DATA SET REQUIRED

EXPLANATION: SORTWKxx data set(s) are required for one of
the following conditions in this execution of MFX: (1)
INCORE=OFF is specified as a PARM, (2) exit E14 or E16 is
activated, (3) the SUM control statement is used, (4) the
OUTREC control statement is used, (5) the checkpoint-restart
facility is used, (6) SORTOUT is a VSAM data set, (7) the
OUTFIL control statement is used. (All conditions only apply to
sort applications.)

WER184A INVALID RETURN CODE FROM E32

EXPLANATION: The return code from merge exit E32 must be 8,
12, or 16.

WER185I SORTIN DCB BLKSIZE GREATER THAN ACTUAL

BLKSIZE, I/O INEFFICIENT
EXPLANATION: The I/O rate is reduced to an inefficient level
because the blocksize specified for the input data set is larger
than the actual blocksize, causing excessive error correction.
ACTION: Correct the blocksize specification for future jobs.

WER186I SVC {nnn,109-rrr} IS INCORRECT VERSION OR NOT A

SYNCSORT SVC - SVC NOT USED - INEFFICIENT SORT
EXPLANATION: The SVC nnn or SVC 109 with router code rrr
was specified as the MFX SVC. The SVC did not return a code
indicating it was at the correct version level, therefore it was not
used. The SVC is either at the wrong MFX release/maintenance
level or is not an MFX SVC. The problem could cause less
efficient I/O and/or loss of SMF records.
ACTION: Notify your system programmer, who should check that
the SVC has been installed in the system libraries, has been
IPLed into the system, was specified via SYNCMAC, and was not
incorrectly overridden via $ORTPARM or the PARM field.

Syncsort MFX Programmer’s Guide 17–21

 WER187A SORTOF CINV SIZE LESS THAN RECORD LENGTH BUT
SPANNING NOT SPECIFIED
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The record length is greater than the control interval
size specified in the definition of the indicated VSAM data set,
but the data set definition did not also include a specification for
spanned records.

WER188A SORTIN IS DIRECT ACCESS/DSCB NOT FOUND/OBTAIN

FAILED
EXPLANATION: The ddname is SORTIN, SORTINnn or
SORTMInn. MFX was unable to successfully issue an OBTAIN
for the specified direct access data set and was therefore unable
to determine the DCB characteristics for the file. The OBTAIN
failed either because the volume parameter was incorrectly
specified for the input file indicated or because the data set was
deleted from the volume. (NOTE: the data set may still be in the
master catalog even though the data set is no longer on the
volume.)

WER189A ddname DCB RECFM REQUIRED

EXPLANATION: The RECFM was not specified on the indicated
input DD statement, nor was it available in the DSCB on disk nor
the tape label, and the TYPE operand was not specified on the
RECORD control statement.

WER190A ddname DUPLICATE OUTFIL SPECIFICATION

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The indicated output file is referred to more than once
in FILES parameters on the OUTFIL control statement.

WER191A ddname BLKSIZE/LRECL INVALID

EXPLANATION: This message is displayed in conjunction with
either WER108I or WER109I which will indicate the invalid DCB
characteristic specification of the input ddname. BLKSIZE and
LRECL must be equal if RECFM=F. BLKSIZE must be evenly
divisible by LRECL if RECFM=FB. BLKSIZE must be greater
than or equal to LRECL + 4 if RECFM=V.

WER192A ddname DCB LRECL MISSING

EXPLANATION: The LRECL was not specified on the indicated
input DD statement, in the DSCB on the disk, in the tape label,
or on the RECORD control statement.

17–22 Syncsort MFX Programmer’s Guide

 WER193A ddname DCB LRECL AND BLKSIZE MISSING
EXPLANATION: The BLKSIZE or LRECL must be specified
either on the indicated input DD statement, in the DSCB on the
disk, or in the tape label. Alternatively, an l1 specification may be
included on the RECORD control statement. None of these
specifications were made.

WER194A SORTOUT DCB REQUIRED/TAPE NOT SL

EXPLANATION: DISP=OLD was specified on the SORTOUT DD
statement, the tape label was not specified as SL in the LABEL
parameter, and required DCB information (LRECL, RECFM,
BLKSIZE) was not specified.

WER195A ddname DCB REQUIRED/VSAM INPUT

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The indicated output file requires additional DCB
information (RECFM, LRECL or BLKSIZE) on its DD statement.

WER196A SORTIN RECFM=VB, LRECL GREATER THAN BLKSIZE

EXPLANATION: RECFM=VB requires the BLKSIZE of the
input ddname to be greater than or equal to LRECL + 4.

WER197A ddname RECFM=F/FB, LRECL/BLKSIZE INVALID

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. BLKSIZE and LRECL were not equal on the
indicated DD statement for RECFM=F, or BLKSIZE was not a
multiple of LRECL for RECFM=FB.

WER198A SORTxx VARIABLE LRECL LESS THAN OR EQUAL TO 4

EXPLANATION: The LRECL specification on the indicated input
or output DD statement did not allow 4 bytes for the RDW plus 1
byte for data. xx could be ‘IN’ or ‘OF’.

WER199A SORTxx RECORD TYPE=V, BLKSIZE LESS THAN OR

EQUAL TO 8
EXPLANATION: The BLKSIZE specified for the indicated input
or output DD statement did not allow 4 bytes for the BDW, 4
bytes for the RDW plus 1 byte of data. xx could be ‘IN’ or ‘OF’.

WER200A ddname RECFM=V/VB LRECL/BLKSIZE INVALID

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. RECFM=V or VB requires the BLKSIZE to be greater
than or equal to LRECL + 4.

Syncsort MFX Programmer’s Guide 17–23

 WER201A ddname IS DIRECT ACCESS/DSCB NOT FOUND/OBTAIN
FAILED
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. MFX was unable to successfully issue an OBTAIN for
the specified direct access data set, and was therefore unable to
determine the DCB characteristics of the indicated file. The
OBTAIN failed either because the volume parameter was
incorrectly specified for the indicated output file, or because the
data set was deleted from the volume. (NOTE: the data set name
may still be in the master catalog even though the data set is no
longer on the volume.)

WER202A ddname RECFM INCOMPATIBLE

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The record format of the output file is not the same as
the input file, the record format of a record provided by an E15, or
the record format created by a JOIN REFORMAT statement.
(Both formats must be either fixed-length or variable-length.) If
you want to convert a variable-length input file into a fixed-
length output file, use the CONVERT parameter of the OUTFIL
or OUTREC control statements. If you want to convert a fixed-
length input file into a variable-length output file, use the FTOV
parameter of the OUTFIL control statement.

WER206A INVALID SVC NUMBER

EXPLANATION: MFX’s processing requires its SVC, but no SVC
number was specified at installation time.
ACTION: Inform your systems programmer of this error
condition.

WER207I SORTCKPT DD STATEMENT MISSING OR INVALID

EXPLANATION: MFX could not take checkpoints because a
SORTCKPT DD statement was not supplied or the statement
specified an invalid device for a checkpoint data set. Invalid
devices include DUMMY data sets or devices other than disk or
tape. Processing continued but checkpoints were not taken.

WER208I MIXTURE OF SORTWK DEVICES

EXPLANATION: SORTWKxx data sets were assigned to
different device types.

17–24 Syncsort MFX Programmer’s Guide

 WER209B xxx PRIMARY AND yyy SECONDARY SORTOUT TRACKS
ALLOCATED, zzz USED
EXPLANATION: It was necessary for MFX to request one or
more secondary allocations for SORTOUT. xxx is the number of
tracks that were initially allocated, yyy is the total number of
tracks acquired via secondary allocation, and zzz is the total
number of tracks actually required to contain the SORTOUT
data set.

WER210I E15 RETURN CODE INVALID, IGNORED

EXPLANATION: A return code of 0 or 4 was passed by an E15
exit routine at a time when these return codes are invalid
because MFX has not passed the E15 a record address. The
invalid return code was ignored by MFX, and a return code of 8
was presumed.

WER211B/I [ ] CALLED BY SYNCSORT; RC=xxxx

EXPLANATION: The sort statistics routine (the name inserted
in the message) is called by MFX. RC gives the code returned to
MFX by the statistics routine. Note that RC=36 is generally
issued when the MFX SVC is not active; the SVC must be
installed to create SMF records. If RC does not equal zero or 36,
see “Before Contacting Precisely Support” on page 18-5.

WER213A INVALID SUM DATA FIELD

EXPLANATION: A field with an invalid data length was
specified on the SUM statement.
ACTION: Correct the field length.

WER215A [SORTOFnn] {INREC,OUTREC} ARITHMETIC

OVERFLOW
EXPLANATION: When using either INREC, OUTREC or
OUTFIL OUTREC, an arithmetic calculation or a data format
conversion had an overflow. An arithmetic calculation overflow
will occur if any intermediate result exceeds 31 decimal digits or
if division by zero is attempted. Overflow may also occur when
converting a number with a value of 4G or more to a 4-byte BI
format or a number with an absolute value of 2G or more to a 4-
byte FI format. An 8-byte BI value is limited to
18446744073709551615. The absolute value of an 8-byte FI or FL
number is limited to 9223372036854775807.
ACTION: Review the arithmetic calculations specified in the
indicated statement for errors. If they appear to be correct,
consider whether the data could possibly cause an overflow or
division by zero. If possible, eliminate any data with questionable

Syncsort MFX Programmer’s Guide 17–25

 values via INCLUDE/OMIT. Consider changing the order of the
calculations to prevent intermediate calculation overflow.

WER216A SUM FIELD OUTSIDE RANGE

EXPLANATION: A sum field on the SUM control statement is
located beyond the record length.

WER217A DYNALLOC {UNIT,STORCLAS} ASSIGNMENT ERROR

EXPLANATION: Either the unit name or storage class name
(DFSMS STORCLAS) is missing or specified incorrectly.

WER219A DYNALLOC FAILED RC=(nnnn) - uuuuuuuu [-SMS

RC=ssss]
EXPLANATION: The execution of the DYNALLOC macro
instruction failed. nnnn represents the error reason code,
uuuuuuuu represents either the unit name or storage class
name, and ssss represents the SMS return code (only present for
certain failures detected by SMS). Two possible reason codes are:
021C - Undefined unit name.
0214 - Unit not available. If all specified units are unavailable
when DYNALLOC is issued, the DYNALLOC request fails.
For other reason codes, see IBM publication z/OS MVS Program-
ming: Authorized Assembler Services Guide SA22-7608.

WER219I DYNALLOC FAILED RC=(nnnn) - uuuuuuuu [-SMS

RC=ssss] SORT PROCESSING CONTINUES
EXPLANATION: Dynamic allocation was unsuccessful. nnnn
represents the error reason code, uuuuuuuu represents either the
unit name or storage class name, and ssss represents the SMS
return code (only present for certain failures detected by SMS).
Sort processing continues with previously allocated SORTWKs
and JCL-allocated SORTWKs. For an explanation of the error
reason code, see IBM publication z/OS MVS Programming:
Authorized Assembler Services Guide SA22-7608.

WER220A INVALID OVERLAPPING OF SUM FIELDS

EXPLANATION: A SUM field overlaps another SUM field, a
SORT/MERGE control field or the Record Descriptor Word of a
variable-length record. All of these are invalid.

WER223A ddname ASCII XLATION, BUT VOLUME IS NOT ASCII

TAPE OR RECFM IS V
EXPLANATION: RECFM=D was specified for the indicated
input or output file which is not a tape data set. (RECFM=D is

17–26 Syncsort MFX Programmer’s Guide

 valid for tape data sets only.) Or, RECFM=D was specified for the
input data set and no DCB was specified for the output data set.
ACTION: In the first case, code correct RECFM for the data set
specified; in the latter case, code DCB characteristics for the
output data set, and rerun the job.

WER224A ddname NOT DEFINED

EXPLANATION: A required DD statement could not be found.

WER225I E35 RC INVALID, IGNORED

EXPLANATION: An invalid return code was received from an
E35 exit routine. If an output data set was not present, the
invalid code was other than 4 or 8, and MFX assumed return code
4. If end of file was reached, the invalid code was other than 8 or
12, and MFX assumed return code 8.

WER227A ddname BLKSIZE GREATER THAN ASCII LIMIT

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The DD statement for an output data set targeted to
an ASCII-labeled tape requested a blocksize greater than 2048
bytes; that violates the standard and cannot be done.

WER228A ddname DCB BLKSIZE GREATER THAN TRACK

CAPACITY
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The BLKSIZE for the indicated output file was
greater than the track capacity of the output device.
ACTION: Specifying the track-overflow RECFM in the DCB may
possibly correct the error condition, or the BLKSIZE should be
reduced.

WER229A ddname DSORG NOT PS/PO

EXPLANATION: The file defined by ddname must be a
sequential data set (PS) or a partitioned data set (PO) member.

WER230A [ddname] xxxxxxxx FIELD OUTSIDE RANGE

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The reason for this may be any of the following:

• A field specified in the SORT/MERGE or JOINKEYS statement is

not located within the first 32750 bytes of the variable-length
record. (This limit is lower if AC, AQ, E, PD0, Y2x or LOCALE CH
fields are used.)

Syncsort MFX Programmer’s Guide 17–27

 • A field specified for INREC, OUTREC, OUTFIL OUTREC,
REFORMAT, SECTION control, (SUB)TOTAL, (SUB)MIN,
(SUB)MAX, (SUB)AVG or HEADER/TRAILER data field is located
beyond the maximum record length.

• INREC, OUTREC, OUTFIL OUTREC, REFORMAT, OUTFIL

IFTRAIL, or HEADER/TRAILER n/col/date/page attempted to build
a record larger than the allowable maximum.

• A REFORMAT statement referenced a join input file field, but

records from that input file were excluded by specifying ONLY on
the JOIN statement.

WER231A [ddname] {INREC,OUTREC} - INVALID DATA FIELD

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. An error was found in the INREC, OUTREC or
OUTFIL OUTREC specification.
ACTION: Check the statement for alphabetic data in a numeric
field, for a parameter value of 0, for an omitted value, for a space
value greater than 256X, for incorrect boundary alignment, and
for inclusion of the "variable portion" of fixed-length input records
in the output records. Also, LINES=ANSI or LINES=(ANSI,n)
may not be used on the OUTFIL statement when using multiline
OUTREC.

WER232A ddname RECFM=VBS, LRECL MISSING

EXPLANATION: A RECFM of VBS was specified for the input
ddname without an accompanying LRECL specification.

WER233A VIO INVALID FOR DYNALLOC

EXPLANATION: VIO is not permitted as a unit device for
dynamic allocation. This is due to a possible performance
degradation if VIO data sets are used as SORTWK.

WER235A [ddname] {INREC,OUTREC,REFORMAT} RDW NOT

INCLUDED
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. Four bytes must be provided for the RDW of the
variable-length output record in the FIELDS parameter of the
INREC, OUTREC, OUTFIL OUTREC, or REFORMAT
specification. These bytes must appear at the beginning of the
record and must not be edited. For REFORMAT, the RDW must
be specified as coming from a variable-length join input data set.

17–28 Syncsort MFX Programmer’s Guide

 WER236A [ddname] {INREC,OUTREC} NULL RECORD
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. A variable-length INREC, OUTREC or OUTFIL
OUTREC output record must contain at least one other data field
in addition to the RDW. Or if multiline OUTFIL OUTREC is
being used, at least one non-blank line must be defined.

WER237I OUTREC RECORD LENGTH=xxxx

EXPLANATION: The xxxx represents the length of the record
after OUTREC processing. OUTREC occurs prior to E35 and/or
SORTOUT/OUTFIL processing. If the data consists of variable-
length records, xxxx represents the maximum record length.

WER238I POTENTIALLY INEFFICIENT USE OF INREC

EXPLANATION: The INREC control statement has been used to
increase the input record length. This can reduce MFX’s
performance because a larger volume of data is being processed
than if the OUTREC control statement were used to perform the
same function. Typically, increasing the record length with
INREC is only useful when expanding SUM fields with leading
zeros to prevent an overflow condition during SUM.
ACTION: Revise the application so that addition of data is
performed in an OUTREC statement. Be sure to adjust the
FIELDS of the SORT, MERGE or SUM control statements if
necessary.

WER239A TYPE PARAMETER REQUIRED

EXPLANATION: There was a VSAM input or output file but the
TYPE parameter was not specified. Or, an E15 or E32 exit
routine is passing all of the records to the sort/merge (no
SORTIN/SORTINnn/SORTMInn), but the TYPE parameter was
not specified on the RECORD control statement.

WER240A ddname UNSUPPORTED DCB FUNCTION

EXPLANATION: The DD statement specified or implied an
attribute which is not supported, e.g., hardware keys for a disk
output data set or a block prefix length other than 0, 4 or L for an
ASCII tape output data set.

WER243I SHORT RECORD FOR SUM

EXPLANATION: One or more variable-length records were too
short to contain all the sum fields specified on the SUM control
statement. These records were therefore not summed. Program
HISTOGRM may be used to determine the length of the shortest
record in the input file.

Syncsort MFX Programmer’s Guide 17–29

 WER244A [ddname] {INREC,OUTREC} SHORT RECORD
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. A variable-length record was too short to contain all
the fields specified on the control statement. Program
HISTOGRM may be used to determine the length of the shortest
record in the input file.

WER246I FILESIZE x
EXPLANATION: The number of bytes of input data sorted or
copied by MFX is given for FILESIZE. This number reflects input
data set, E15, JOIN, INCLUDE/OMIT, and INREC processing.
Note the following:

• For MAXSORT, the FILESIZE is given in kilobytes for each

individual sort in a WER351I message; the FILESIZE in the
WER246I for the final merge is the sum of the individual sorts’ sizes
and, because of truncation in each intermediate sort, may not be
exact.

• For variable-length record cases, when WER246I is issued and

WER054I is also issued, the number of bytes processed includes a
single record descriptor word for each record. When WER246I is
issued and WER054I is not issued, the number of bytes processed
includes the record descriptor words for all segments of any multi-
segment records.

WER247A ddname HAS INCOMPATIBLE LRECL

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. There is a conflict between the LRECL specification
for the indicated output file and either the post-OUTFIL or post-
OUTREC record length. Padding of records is not permitted after
OUTFIL processing, so the LRECL may not be greater than the
post-OUTFIL record length. Alternately, truncation of records is
not permitted after the OUTREC statement or the OUTFIL
OUTREC processing, so the LRECL may not be less than the
post-OUTREC record length.

WER250A [ddname] INCLUDE/OMIT FIELD BEYOND RECORD

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. A compare field specified for an INCLUDE/OMIT/
WHEN/BEGIN/END/TRLID comparison extended beyond the
end of the record. If it is expected that the input record (or for
OUTFIL, the record after INREC or OUTREC control statement
processing) will not be long enough to contain all the INCLUDE/

17–30 Syncsort MFX Programmer’s Guide

 OMIT fields, refer to “VLTESTI” on page 5-32 for alternate
methods of handling this short record condition.

WER251A INCLUDE/OMIT INVALID yyyyyyyyyy

EXPLANATION: The invalid relational condition represented by
yyyyyyyyyy was found in the INCLUDE/OMIT/WHEN/BEGIN/
END/TRLID parameter specification.

WER253A INCLUDE/OMIT FORMATS INCOMPATIBLE

EXPLANATION: A relational condition specified in an
INCLUDE/OMIT/WHEN/BEGIN/END/TRLID comparison
contains an invalid field-to-field, field-to-constant or field-to-
mask comparison. Note that if LOCALE processing has been
specified, a CH to BI comparison is not supported.

WER254A ddname VSAM {OPEN,CLOSE} ERROR - xx

EXPLANATION: An error occurred during an attempt to OPEN
or CLOSE a VSAM file defined by ddname. For the definition of
the error number, xx, consult the following IBM publication:
• DFSMS Macro Instructions for Data Sets

Note: If xx is A0, it is likely that the VSAM data set is empty, and that
VSAMEMT=NO is in effect. Pass the parameter VSAMEMT=YES as a
possible solution.

WER255A VSAM LOGICAL ERROR xx ON {INPUT,OUTPUT}

EXPLANATION: An error occurred while processing a VSAM
data set. For the definition of the hexadecimal error number
represented by xx, see the following IBM publication:
• DFSMS Macro Instructions for Data Sets

Note: If xx is 0C or 08 on output, it is likely that the VSAM output data

set was created with REUSE, the VSAM data set is not empty, and
RESET is not in effect. Pass the parameter RESET as a possible solu-
tion.

WER256I ddname VSAM file, RECORDS PADDED ON OUTPUT

EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The fixed-length VSAM LRECL for the indicated
output file is greater than the length of the records at the end of
MFX processing. MFX padded the output records with filler
characters on the right.

Syncsort MFX Programmer’s Guide 17–31

 WER257I INREC RECORD LENGTH=xxxxx
EXPLANATION: xxxxx represents the length of the record
immediately after INREC processing. If you have variable-length
records, xxxxx represents the maximum record length.

WER258A DUPLICATE DDNAME: SORTINxx

EXPLANATION: Two input files for a merge have the same
number. The file number is given.

WER259A DUPLICATE ALTSEQ STATEMENT

EXPLANATION: Two ALTSEQ control statements were found.

WER260I RECOVERY FROM B37 SUCCESSFUL. SORT

PROCESSING CONTINUES
EXPLANATION: MFX recovered from a B37 abend and
continued processing.

WER262I REENTRANT SORT NOT RESIDENT - INEFFICIENT

SORT
EXPLANATION: The resident MFX load module(s) were loaded
into the private area instead of being executed from the Link
Pack Area/Extended Link Pack Area. This situation may have
occurred because the module(s) were found in a STEPLIB/
JOBLIB DD data set. Loading the resident modules into the
private area limits the amount of virtual storage available to the
sort and may reduce the efficiency of the sort.
ACTION: Contact the systems programmer in charge of MFX
installation.

WER263A MULTIVOLUME SORTWK DATA SETS ARE NOT

SUPPORTED
EXPLANATION: MFX does not support the use of multi-volume
disk SORTWK data sets. (However, if MFX only requires the use
of the space on the first volume of a multi-volume SORTWK file,
this error message will not be issued.)
ACTION: Remove the volume count subparameters of the UNIT
parameter on all SORTWK DD statements that specify more
than one volume.

WER264A UNEQUAL RECORD LENGTHS - VSAM SORTxx - TYPE=F

EXPLANATION: A record in a fixed-length VSAM input data set
was encountered whose length was not equal to the length
specified in the RECORD statement or VSAM cluster definition.
xx could be ‘IN’ or ‘MI’.

17–32 Syncsort MFX Programmer’s Guide

 ACTION: Use the IDCAMS utility to identify and correct the
records in error.

WER265A ddname VSAM CONCATENATED INPUT NOT ALLOWED

EXPLANATION: The ddname indicated represents an input file
which consists of concatenated VSAM data sets. MFX does not
support concatenated VSAM input files.
ACTION: MFX is able to read multiple VSAM and non-VSAM
input files through the MULTIIN facility. See “Chapter 12
Multiple Input Files” for information on how to use MULTIIN.

WER266A ALTPARM - PARM LENGTH GT MAX SUPPORTED

EXPLANATION: The length of the parameter list passed
through the alternate parameter data set exceeded the 256 byte
limitation.

WER267A statement STATEMENT: STATEMENT NOT FOUND

EXPLANATION: A required SORT/MERGE or RECORD
statement (as indicated in the message text) is missing.

WER268A statement STATEMENT: SYNTAX ERROR

EXPLANATION: An MFX control statement, as indicated in the
message text, contains a syntax error. The next line will contain
an '*' indicating the approximate location of the syntax error.

WER269A statement STATEMENT: DUPLICATE STATEMENT

FOUND
EXPLANATION: More than one ALTSEQ, DUPKEYS, END,
JOIN, JOINKEYS, INCLUDE/OMIT, INREC, MODS, OUTREC,
RECORD, REFORMAT, SORT/MERGE, or SUM statement was
found, as indicated.

WER270A statement STATEMENT: DUPLICATE PARM FOUND

EXPLANATION: A single parameter was multiply specified on
the indicated MFX control statement; or a single parameter was
specified both in the invoking parameter list and in the control
statements.

WER271A statement STATEMENT: NUMERIC FIELD ERROR

EXPLANATION: A numeric field has been improperly specified
on the indicated MFX control statement.

WER272A statement STATEMENT: PARMS NOT FOUND

EXPLANATION: Required parameters have not been included on
the indicated MFX control statement.

Syncsort MFX Programmer’s Guide 17–33

 WER273A BLANK STATEMENT FOUND
EXPLANATION: A blank statement has been encountered.

WER274A CONTINUATION STATEMENT ERROR FOUND

EXPLANATION: MFX has encountered a statement containing a
continuation indicator, but cannot locate a continuation
statement which should follow.

WER275A NO KEYWORDS FOUND ON CONTROL STATEMENT

EXPLANATION: A required keyword has not been specified on
an MFX control statement.

WER276B SYSDIAG=nnnnnnnn,nnnnnnnn,nnnnnnnn,nnnnnnnn
EXPLANATION: This message contains internal diagnostic
information intended for use by Precisely Support.

WER277A [ddname] {INREC,OUTREC} - INVALID USE OF VL

(VARIABLELENGTH OUTPUT)
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The VL subparameter was used within JFY/SQZ and
is invalid for one of the following reasons:

• JFY/SQZ is used within the OVERLAY parameter.

• An output record field with a starting column is specified after JFY/

SQZ.

• The LENGTH subparameter is used within the JFY/SQZ.

• The input record to INREC/OUTREC is fixed-length (unless used on

an OUTFIL statement where the FTOV parameter is specified).

• JFY/SQZ is used within an OUTFIL IFTHEN clause with either

WHEN=INIT or HIT=NEXT, and the FTOV parameter is specified.

WER292A BLOCK E15/E35 ERROR, REASON CODE nn

EXPLANATION: A Block Exit error was found in the E15 or E35
routine module, where nn represents the error reason code.
• 01 Block List Parameter Area or Block List address Invalid
• 02 Block List type Invalid
• 03 Block List boundaries Invalid
• 04 Block address (Type 2) or Record address (Type 1) Invalid
• 05 Block Exit Return Code Invalid

17–34 Syncsort MFX Programmer’s Guide

 • 06 Record Length Invalid (Variable length Record)

WER300A SORTBKPT DD STATEMENT REQUIRED

EXPLANATION: The SORTBKPT DD statement was not
included in the job stream. This is a required data set for all
MAXSORTs.

WER301A SORTBKPT DATA MUST RESIDE ON DISK

EXPLANATION: The SORTBKPT data set must be allocated to a
disk device.

WER302A SORTBKPT TRACK CAPACITY TOO SMALL

EXPLANATION: Direct access devices with a track capacity
smaller than 3600 bytes cannot be used for the SORTBKPT data
set.

WER303A SORTBKPT SYSTEM OPEN FAILURE

EXPLANATION: The operating system could not open the
SORTBKPT data set.
ACTION: Check to see that the DD statement is correct.
Determine if operating system is at proper maintenance level.

WER304A SORTBKPT RECORD FORMAT ERROR

EXPLANATION: There is a record format error in the
SORTBKPT data set.
ACTION: Check that the SORTBKPT DD statement points to the
correct DSNAME. Check that the data set has not been
inadvertently written into and modified. Use the HEX function
on the OUTREC statement or OUTREC parameter on the
OUTFIL statement to get a hex format listing of the data.

WER305A SORTBKPT RECORD EXCEEDS BLKSIZE

EXPLANATION: The use of an excessive number of parameters
in a control statement has caused the SORTBKPT data set to
overflow the maximum blocksize limit of 32760.
ACTION: Reduce the size of the control statement specification if
possible, or convert the application from a MAXSORT to a
conventional sort.

WER306A RESTART FROM BREAKPOINT PROHIBITED

EXPLANATION: The SORTBKPT data set indicates that a
program-initiated sort or a sort with exit programs tried to
restart from a breakpoint.
ACTION: Use z/OS checkpoint facilities since only these will save
your work areas and the program memory for restart.

Syncsort MFX Programmer’s Guide 17–35

 WER307A SORTBKPT RECORD SEQUENCE ERROR
EXPLANATION: An out-of-sequence record was read from the
SORTBKPT data set.
ACTION: Use the HEX function on the OUTREC statement or
OUTREC parameter on the OUTFIL statement to get a
hexadecimal listing of the data set for analysis. See if the data set
was damaged by another program. Check system for hardware
error.

WER308A BREAKPOINT ID NOT FOUND ON SORTBKPT

EXPLANATION: The parameter RESTART=id was specified but
id could not be found.
ACTION: Check spelling, correct, and return.

WER309A SORTOUXX DATA MUST BE ON DISK OR TAPE

EXPLANATION: Intermediate sort output data was allocated to
an unsupported device. Only disk or tape is allowed.
ACTION: Allocate SORTOUxx data to either disk or tape.

WER310A SORTOUXX DEVICE MIXING PROHIBITED

EXPLANATION: Intermediate sort output was allocated to both
tape and disk in the same job or to a mixture of disk device types.
ACTION: Allocate all intermediate sort data to the same device
type.

WER311A DISK SORTOUXX REQUIRES SORTOUXX DD

EXPLANATION: No SORTOUxx DD statements were found so
there was no place to store intermediate sort output.
ACTION: Supply one or more SORTOUxx DD statements with xx
represented by 01 to 99.

WER312A TAPE SORTOUXX REQUIRES SORTOU00 DD

EXPLANATION: One or more SORTOUxx DD statements were
allocated to tape but the SORTOU00 statement was not present.
ACTION: Allocate a tape unit using the SORTOU00 DD
statement.

WER313A SORTOUXX DEVICE NOT SUPPORTED

EXPLANATION: The SORTOUxx DD statements specify an
unsupported device type.
ACTION: Change the device allocation of the SORTOUxx data
set.

17–36 Syncsort MFX Programmer’s Guide

 WER314A INSUFFICIENT VIRTUAL STORAGE FOR MAXSORT
EXPLANATION: MAXSORT cannot run efficiently in the
amount of virtual storage provided.
ACTION: Increase virtual storage or decrease the number of tape
units requested by MINMERGE.

WER315A SORTOUXX BLKSIZE GT TRACK CAPACITY

EXPLANATION: Intermediate sort output is on disk, but the
input data set requires too large a blocksize for a disk device.
ACTION: Allocate intermediate sort output to tape and rerun.

WER316A INSUFFICIENT SORTOUXX DD STATEMENTS

EXPLANATION: The data to be sorted requires one or more
additional data sets.
ACTION: Recalculate and restart the job including additional
SORTOUxx DD statements. (Make sure each statement’s
number is greater than the last one you put in.)

WER317I MAXSORT OPTION SELECTED

EXPLANATION: A MAXSORT was requested.

WER318I INPUT CARDS IGNORED - SORTBKPT USED

EXPLANATION: The control statement just listed on SYSOUT
for a breakpoint/restart were not used to control sorting.
Whatever control statements were specified when the job was
started were used. (They may be the same as the statements just
listed, however.)

WER319I SORT RESTARTED AT BKPT xxxxxxxxxxxx

EXPLANATION: This message identifies the breakpoint id from
which MAXSORT resumes execution on a breakpoint/restart.

WER320I INEFFICIENT SORTOUXX BLKSIZE FORCED

EXPLANATION: Due to the constraints between the amount of
memory and the value specified for MAXMERGE, MAXSORT
was forced to compromise and choose a smaller blocksize than
would permit efficient buffering in sorts and merges.
ACTION: If you wish a more efficient MAXSORT, either increase
the amount of memory or reduce the number specified for
MAXMERGE. This will permit a larger blocksize to be chosen
which will allow multiple buffering of all the intermediate sort
output data.

Syncsort MFX Programmer’s Guide 17–37

 WER321B SORTOUXX BLKSIZE=xxxxx
EXPLANATION: This gives the blocksize that MAXSORT has
chosen for intermediate sort output.

WER322A TAPE DYNALLOC FAILURE - CODE=xxxx

EXPLANATION: Attempts to dynamically allocate tape units for
a merge phase met with unexpected failure. Code xxxx gives the
hexadecimal return code from the dynamic allocation request.
For an explanation of this code, see IBM publication z/OS MVS
Programming: Authorized Assembler Services Guide SA22-7608.

WER323A BKPT DATA AT DIFFERENT RELEASE LEVEL

EXPLANATION: The SORTBKPT data was created by a
different MFX release than the MFX program reading it. Because
of this, the breakpoint data cannot be processed.
ACTION: Restart this job and run under the same MFX release
that you started with.

WER324A TAPENAME CLASS NOT FOUND ON SYSTEM

EXPLANATION: The tapes could not be dynamically allocated
because a TAPENAME was specified that was not generated into
the operating system.
ACTION: Check with the systems programmer for acceptable
unit names.

WER325A MAXSORT STOPPED BY OPERATOR

EXPLANATION: The operator responded to a message by
stopping the sort. The sort may be restarted from the last
breakpoint or checkpoint.

WER326A DYNALLOC UNALLOC FAILURE - CODE=xxxx

EXPLANATION: Attempts to dynamically deallocate tape units
met with unexpected failure. Code xxxx gives the hexadecimal
return code from the dynamic deallocation request. For an
explanation of this code, see IBM publication z/OS MVS
Programming: Authorized Assembler Services Guide SA22-7608.

WER327A INSUFFICIENT UNITS FOR MINIMAL MERGE

EXPLANATION: Too few tape units were allocated to meet the
number specified in MINMERGE. Either too few SORTOUxx DD
were supplied or the z/OS system was unable to dynamically
allocate enough units.
ACTION: Restart the job with additional SORTOUxx DD
statements.

17–38 Syncsort MFX Programmer’s Guide

 WER328A SORTOUXX SYSTEM OPEN FAILURE
EXPLANATION: The operating system could not open the
SORTOUxx data sets.
ACTION: Check to see that SORTOUxx DD statements are
correct. Determine if operating system is at a proper
maintenance level.

WER329A SORTOU00 SYSTEM RDJFCB FAILURE

EXPLANATION: The operating system could not read the Job
File Control Block for MFX analysis.
ACTION: Determine if operating system is at a proper
maintenance level.

WER330A SPECIFIED SORTING TIME HAS EXPIRED

EXPLANATION: The time limit specified in the SORTTIME
parameter has expired. The job may be restarted from the last
breakpoint or checkpoint.

WER331A SYSTEM CHECKPOINT FAILURE

EXPLANATION: Request for z/OS checkpoint facilities failed.
ACTION: Ascertain that the SORTCKPT DD statement was
correctly specified. Check that rules for the use of checkpoint
were not violated.

WER332A TOO MANY INTERMEDIATE SORTS - INCREASE

SORTWORK SPACE
EXPLANATION: Only 99 intermediate sorts are allowed in a
MAXSORT application.
ACTION: Increase the SORTWORK space available to
MAXSORT so that each intermediate sort will process more data,
reducing the number of intermediate sorts required. Ensure that
the MINWKSP and MAXWKSP values are sufficient to allow
additional space to be acquired. The application does not have to
be restarted from the beginning. If a MAXSORT breakpoint/
restart is allowed in the application, restart from an earlier
breakpoint with a sufficient amount of SORTWORK space
available to process the file within the 99 intermediate sort limit

WER333I zHPF USED FOR {INPUT, OUTPUT}

EXPLANATION: MFX used zHPF channel programs to improve
the performance of the input and/or output phase.

Syncsort MFX Programmer’s Guide 17–39

 WER350I {SORT/MERGE} # XX COMPLETE {AT BREAKPOINT/AT
CHECKPOINT} bbbbbbbbbbbb, DIAG=hhhh,hhhh...
EXPLANATION: This message tells which individual sort or
merge has completed. Restart can be performed from the
breakpoint or checkpoint id given in bbbbbbbbbbbb. If restart is
not possible the above message will read:
SORT/MERGE # XX COMPLETE.
The hexadecimal information following the DIAG keyword is
likely to change from execution to execution. It is internal
diagnostic information intended for use by Precisely Support.

WER351I DATA SIZE xxxx KB [FROM yy WAY MERGE]

EXPLANATION: The amount of data that was processed for the
current MFX individual sort/merge is given in kilobytes. When a
merge is processed yy gives the number of tape units used.

WER352I DYNAMICALLY ALLOCATED TAPE UNITS - XX

EXPLANATION: The number of tapes drives that were
dynamically allocated for the current merge pass is given.

WER353I STARTING TIME hh.mm.ss - ENDING TIME hh.mm.ss

EXPLANATION: The starting and ending times in hours,
minutes, and seconds of the individual sort or merge just
completed are given.

WER354I ----------------------DATA SET STATUS----------------------

EXPLANATION: This is a header. Messages relating to data sets
will follow.

WER355I {DSN=dsname/VOL SERS = vvvvvv...}

EXPLANATION: The data set names of the tapes for
intermediate sort output are given. The tape volumes are listed
for tape intermediate sort output. Retain these reels for input to
a later merge.

WER356I SORTOUXX DD STATEMENT IS ACTIVE

EXPLANATION: The disk data set allocated to the SORTOUxx
DD statement is needed as input to a subsequent merge. Be sure
to keep it in case restart is necessary.

WER375D jobname.stepname - MAXSORT BKPT id

TIME ESTIMATE: XXX MINUTES UNTIL NEXT
NOTIFICATION.
REPLY 'GO' TO CONTINUE, 'STOP' TO TERMINATE
EXPLANATION: A long-running MAXSORT has exhausted its
assigned block of computer time.

17–40 Syncsort MFX Programmer’s Guide

 ACTION: The operator’s decision should be based on scheduling
priorities and the estimated time of the sort. A 'GO' reply will
permit sort execution to proceed in stages. This message is
generated at discrete intervals so that the operator can again opt
to continue or terminate its execution.

WER376D jobname.stepname - MAXSORT BKPT id

aaa TAPE UNITS ALLOCATED TO jobname
bbb TAPE UNITS NEEDED FOR BEST PERFORMANCE
TIME ESTIMATE USING aaa TAPE UNITS -
xxxx MINUTES TO {NEXT BREAKPOINT | END OF JOB}
REPLY 'GO' TO CONTINUE, 'STOP' TO TERMINATE, 'NN'
# UNITS
EXPLANATION: The first time this message is generated, it
indicates that MAXSORT has dynamically allocated the optimum
number aaa of tape drives up to MAXMERGE. Reissued, this
message documents MAXSORT’s response to the operator’s
previous reply of 'NN' tape units. 'NN' represents the total
number of tapes that will be allocated.
ACTION: Given a reply of 'NN' tape drives, MAXSORT will
attempt to satisfy the operator’s request. For 'NN' larger than
aaa, MAXSORT will try to raise its allocation to 'NN'. (The
operator can delay the request for more tape units in order to give
other jobs time to free any tape drives they are using.) The above
message is reissued and the operator can see how the decision
will affect sort execution.
As soon as allocations and time estimates are satisfactory, the
reply 'GO' will cause continued execution using the allocated tape
units. If allocation or time estimates are not satisfactory, the job
may be terminated (reply 'STOP') or a new number 'NN' of units
may be requested.

WER377D jobname.stepname - MAXSORT BKPT id

INSUFFICIENT TAPE UNITS AVAILABLE
aaa TAPE UNITS ALLOCATED TO jobname
bbb TAPE UNITS NEEDED TO CONTINUE EXECUTION
REPLY 'RETRY' TO GET UNITS, 'STOP' TO TERMINATE
EXPLANATION: MAXSORT cannot immediately acquire enough
tape drives to make continued processing worthwhile.
ACTION: The operator can wait until other tape drives have been
released, then reply 'RETRY'. If enough drives are now available,
execution continues. Otherwise the above message is repeated.
Eventually enough tape drives become available or the operator
terminates the job with a 'STOP' response.

Syncsort MFX Programmer’s Guide 17–41

 WER378I NO ADDITIONAL TAPE UNITS EXIST FOR GENERIC
CLASS tapename
EXPLANATION: All tape units on the system within the
TAPENAME class have been allocated. Further DYNALLOC
attempts will fail to acquire more tape units. Message WER376D
or WER377D will follow.

WER390A MINIMUM SORTWK SPACE NOT AVAILABLE

EXPLANATION: MAXSORT could not obtain enough SORTWK
disk space to run. When MAXSORT is executing with larger
storage values, MFX may need to automatically raise MINWKSP,
overriding the specified MINWKSP value. Therefore, it may
erroneously appear that JCL SORTWKs provided enough space
to satisfy MINWKSP when this message was posted.
ACTION: Correct SORTWK volume, primary, and secondary
allocations. Restart the job.

WER391A INSUFFICIENT VIRTUAL STORAGE FOR SORTBKPT

BUFFER
EXPLANATION: MAXSORT was unable to obtain the necessary
3600-byte buffer space from the operating system.
ACTION: Check to see that sufficient virtual storage was
allocated to the sort.

WER392A SORTBKPT FORMAT ERR - VBS PROCESSING

EXPLANATION: MAXSORT attempted to read back control
information associated with VBS input data and found a format
error in the SORTBKPT data set.
ACTION: Contact Precisely Support or call your MFX support
representative.

WER393I TURNAROUND MAXSORT SORT PERFORMED

EXPLANATION: The amount of input data was small enough to
fit entirely on SORTWK disk space, so sorted data was produced
in one MFX pass.

WER394A SORTOUXX DD STMT REQUIRED FOR MERGE

EXPLANATION: The above DD statement was required for disk
intermediate sort output as input to a merge but could not be
found.
ACTION: Supply the missing DD statement.

17–42 Syncsort MFX Programmer’s Guide

 WER395A INVALID SORTOU00 OR SORTOUxx DSN PREFIX
EXPLANATION: The BKPTDSN parameter was used, but the
required trailing period was not specified as part of the DSN
prefix.
ACTION: Add a trailing period to the parameter specification.

WER396A LKED DD STATEMENT MISSING OR INVALID

EXPLANATION: A MODS statement specified at least one exit
to be link-edited by MFX, but a SYSPRINT and/or SYSLIN and/
or SYSLMOD DD statement is missing. All of these statements
are required for link-editing. Or, the SYSLMOD DD statement
does not refer to a data set on a direct access device.
ACTION: Supply the missing DD statement(s) or adjust the
SYSLMOD DD statement as appropriate.

WER400A ddname IS AN UNINITIALIZED SEQUENTIAL DISK

DATA SET
EXPLANATION: The input data set was allocated but never
opened for output. Therefore, there is no valid data or end-of-file
mark in the data set. This condition usually occurs when a
program abends and the steps to create the data are bypassed.
ACTION: Write the appropriate data or end-of-file mark in the
data set, or see “UNINTDS” on page 5-30.

WER401A CSECT NAME DIFFERENT THAN MEMBER NAME

EXPLANATION: The MODS statement specified an exit routine
module in SYSIN that was not found.
ACTION: Either change the member name in the MODS
statement to match the module name or reassemble the exit
module with a name to match the member name on the MODS
statement.

WER402A SORTMODS STOW FAILURE

EXPLANATION: While copying an exit routine from SYSIN to
SORTMODS, MFX attempted unsuccessfully to store (STOW) the
exit routine in the SORTMODS directory. This condition is
caused either by specifying insufficient directory blocks when
creating the SORTMODS data set or by the presence of a member
or alias with the same name as the exit routine in the
SORTMODS data set, or by a hardware failure.
ACTION: Check the SORTMODS directory names for a member-
name conflict and rerun the job step.

Syncsort MFX Programmer’s Guide 17–43

 WER403A xxxxxxxx NOT VALID FOR MAXSORT
EXPLANATION: xxxxxxxx denotes the feature that is not
supported when using MAXSORT.
ACTION: Either remove the feature specification or convert the
application not to invoke MAXSORT.

WER404I {SORTXSUM,SORTXDUP}: RECFM= ;LRECL= ;

{BLKSIZE=,CISIZE=} [;CINV ACCESS]; RCD OUT n
EXPLANATION: This informational message lists the DCB
characteristics used by MFX to process the SORTXSUM/
SORTXDUP file, as well as the number of records (n) that were
written to the data set. For a VSAM data set, the CISIZE is
provided; if control interval access was used, the CINV ACCESS
portion of the message will be displayed.

WER405I ddname DATA RECORDS OUT n, TOTAL RECORDS OUT

y
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The n represents the number of data records
(exclusive of HEADERS/TRAILERS and multi-record OUTREC)
in each output data set. The y represents the total number of
records in each output data set (data records, HEADERS/
TRAILERS and multi-record OUTREC records). Note that the
total number of lines written to the line printer may be greater
than the actual record count since multiple lines can be
generated from one data record using ANSI control characters.

WER406A ddname HEADER/TRAILER/DATA LINES EXCEED PAGE

SIZE
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The number of lines generated by some HEADER
and/or TRAILER and/or multiline OUTREC parameters is
greater than or equal to the number of lines to be written per
logical page as specified by the LINES parameter. If LINES has
not been coded, this number defaults to 60.
ACTION: Reduce the number of HEADER/TRAILER lines
generated or increase the number of lines in the LINES
parameter so that a minimum of all output lines from 1 data
record can be written per logical page.

WER407I UNUSABLE SORTWK DEVICE ALLOCATED, UNIT=VIO

EXPLANATION: A VIO data set was allocated during dynamic
allocation. The device was held for the duration of the sort;
however, the device was not used for SORTWK storage.

17–44 Syncsort MFX Programmer’s Guide

 ACTION: For future executions, ensure that the DYNALLOC
runtime parameter specifies a unit name that does not cause a
VIO data set to be allocated.

WER409A MOD ON SYSIN NOT FLAGGED AS SYSIN MODULE

EXPLANATION: An object deck was found in the SYSIN data set
that, according to the MODS statement, was not specified as
belonging in SYSIN.

WER410B xxx BYTES OF VIRTUAL STORAGE AVAILABLE ABOVE

THE 16-MEGABYTE LINE, yyy BYTES RESERVE
REQUESTED, zzz BYTES USED
EXPLANATION: The amount of virtual storage above the 16-
megabyte line available (free) when MFX received control is
represented by x’s. The amount of virtual storage that the user
requested MFX to reserve above the 16-megabyte line is
represented by y’s. The amount of virtual storage used by MFX
above the 16-megabyte line is represented by z’s.

WER411B nnn BYTES OF EMERGENCY SPACE ALLOCATED

ABOVE THE 16-MEGABYTE LINE
EXPLANATION: The indicated amount of virtual storage above
the 16-megabyte line has been set aside by MFX for use by other
programs (e.g., program invoking the sort, system SVCs, tape
management system.)

WER412I ERROR TAKING SYSTEM CHECKPOINT. PROCESSING

CONTINUES
EXPLANATION: An error occurred when MFX attempted to take
a user-requested checkpoint. Sort/merge processing continued;
however, a usable checkpoint may not exist. Refer to the IHJxxxx
message in the job log to determine the cause of the error.

WER414A ddname OPEN ERROR ON AN UNINITIALIZED SEQUEN-

TIAL DISK DATA SET
EXPLANATION: An error occurred during an OPEN of a multi-
volume uninitialized sequential disk data set being used for
ddname. When the UNINTDS=YES option has been selected,
either by default or parameter override, MFX will need to open
for output a multi-volume uninitialized disk data set in order to
set the DS1IND80 flag in the format-1 DSCB of the first volume.
Typically this error will occur if the MFX step does not have the
authority to open the data set for output processing.
ACTION: In a separate step prior to the MFX invocation, write
the appropriate end-of-file mark in the first volume of the multi-
volume data set.

Syncsort MFX Programmer’s Guide 17–45

 WER415B DSM FACILITY DISABLED
EXPLANATION: MFX’s dynamic storage management feature
was not active for this sort execution.

17–46 Syncsort MFX Programmer’s Guide

 WER416B

access-method WAS USED FOR ddname 

 
ddname: EXCP'S=eee [,UNIT=uuuu] [,DEV=dddd] [,CHP=cccccccc,n][,VOL=vvvvvv] 
 
TOTAL OF xxx EXCP'S ISSUED FOR totalid 

EXPLANATION: This message provides summary I/O tuning

information for files processed by MFX. The first form is used
when an access method other than EXCP is used for a file. It uses
a generic term for the access method (BSAM, HIPERBATCH,
etc.) and the file for which it was used. When EXCP is used, the
message takes on the second form which has the component parts
listed below. Some of these components may or may not be
included in the message depending on the level of the operating
system and the availability of the information within MFX.
EXCP'S=eee "eee" identifies the number of EXCPs issued
for the file. For input files such as SORTIN,
this is the total EXCPs issued for all concat-
enated input sets.
UNIT=vuuuu "uuuu" is the unit type on which the data
set resides. For files that can consist of con-
catenations or multi-volume data sets, the
unit type displayed is for the first volume of
the first data set.
DEV=dddd "dddd" is the device name for the first or
only device for the file.
CHP=cccccccc,n This field identifies the channel paths
available to the first or only device. ''n'' is
the number of PAV aliases available.
VOL=vvvvvv This field is displayed for only DASD
devices and identifies the volume serial
number of the first or only volume for the
file.
For certain types of sorts, MFX may dynamically allocate data
sets other than SORTWKxx data sets for use in the sorting pro-
cess, and this can occur whether or not normal dynamic alloca-
tion of sortwork data sets is enabled. When used, such data sets
are collectively represented in a single WER416B message using
a ddname of "SORTWK&&" for the purpose of reporting EXCPs
issued against them.

Syncsort MFX Programmer’s Guide 17–47

 In the third form of the message, xxx provides a total of the
EXCPs issued for SORTWORKS, SORTING, COPYING, or
MERGING, as identified by "totalid."

WER417A UNEQUAL MAINTENANCE LEVELS: x,y,z

EXPLANATION: The module x and MFX maintenance levels do
not correspond. y represents the maintenance level of the x
module; z represents the PTF number recorded in the MFX
maintenance record.
ACTION: Contact the systems programmer in charge of MFX
maintenance. This problem is most likely caused by a lack of
synchronization between maintenance levels of modules in the
SYNCLINK, SYNCRENT, and SYNCLPA libraries. Note that the
MFX maintenance record is in the SYNCLINK library.

WER418I DATASPACE(S) AND/OR ZSPACE USED

EXPLANATION: MFX has dynamically chosen to use data space
or ZSPACE during the execution of the sort. ZSPACE is a
technique within MFX created as a replacement for hiperspace. It
allows native use of the central storage resources which are
available. This technique eliminates the additional overhead
produced when hiperspace is simulated by the operating system
in a z/Architecture environment. It provides superior CPU
performance and reduced system overhead compared to a
conventional hiperspace application.

WER420I COBOL ACCELERATOR ACTIVE

EXPLANATION: MFX’s high performance access method was
used for accessing a COBOL file.

WER422A SORTOUT STOW FAILURE

EXPLANATION: When writing to SORTOUT, MFX attempted
unsuccessfully to store (STOW) the SORTOUT PDS member in
the SORTOUT directory. This condition is caused by specifying
insufficient directory blocks when creating the SORTOUT data
set.
ACTION: Recreate the SORTOUT data set with more directory
blocks and rerun the job step.

WER423I DYNAMIC ALLOCATION RETRY - WAITING FOR SPACE

EXPLANATION: The DYNALLOC facility is being used to
acquire sortwork space, but there is currently insufficient disk
space on the system to satisfy the request. MFX will wait the
prescribed number of minutes as specified by the DYNALLOC
option and then retry the request.

17–48 Syncsort MFX Programmer’s Guide

 WER424I DYNAMIC ALLOCATION RETRY SUCCESSFUL
EXPLANATION: The dynamic allocation of sortwork space after
a DYNALLOC RETRY attempt was successful. Sort processing
continues.

WER425A CONVERT FEATURE CANNOT BE USED WITH OVERLAY

OR IFTHEN; OUTFIL CONVERT REQUIRES OUTREC
EXPLANATION: The OUTREC CONVERT feature cannot be
used with OVERLAY or IFTHEN parameters. It also must be
used together with the OUTREC parameter.

WER426I SORT INTERNAL ERROR - RECOVERY ATTEMPT IN

PROGRESS
EXPLANATION: The presence of this message indicates that an
automatic retry of the MFX execution has been initiated. If the
error recovery is successful, the MFX SYSOUT listing will
contain a subsequent set of messages representing the complete
information about the execution. The subsequent set of messages
may be separated from the initial set of listings by a diagnostic
output of significant size. The new listing will contain the
message WER427I.

WER427I RECOVERY ATTEMPT IN PROGRESS

EXPLANATION: The set of SYSOUT messages containing the
WER427I will be from the automatic retry execution. Examine
these messages to insure that it also contains a WER052I
message indicating a successful completion of the MFX
execution. In addition, a successful MFX recovery will complete
with a return code of zero. Even if the WER426I and WER427I
messages are present, this in itself does not constitute a
successful recovery unless zero is returned for the step
completion code.
If an execution of MFX does utilize the recovery facility, whether
successfully or not, the Precisely Support should be contacted so
that the underlying error can be investigated and resolved.

WER428I CALLER-PROVIDED IDENTIFIER IS "xxxx"

EXPLANATION: MFX was invoked by another program, and
that program used a 31-bit or 64-bit parameter list where the
"call identifier" parameter was specified. xxxx is the identifier
specified by the calling program.

WER431I COPY SUBSTITUTED FOR MULTIPLE OUTFILS

EXPLANATION: The SORT or COPY multiple output
application (multiple OUTFILs) has been automatically

Syncsort MFX Programmer’s Guide 17–49

 converted by MFX to a single SORT or COPY operation followed
by one or more COPY operations.
If system resources are available and the output files of a multi-
ple output application have identical specifications, MFX will
make this type of change to take advantage of system resources
to improve the application’s performance.

WER432I {SORT,MERGE} FORMAT OPERAND IGNORED

EXPLANATION: On either a SORT or MERGE control
statement, the format of the keys was specified in both the
FIELDS and FORMAT parameters. MFX ignores the FORMAT
parameter and uses the individual format specifications within
the FIELD parameter.

WER433I SUM FORMAT OPERAND IGNORED

EXPLANATION: On a SUM control statement, the sum field
format was specified in both the FIELDS and FORMAT
parameters. MFX ignores the FORMAT parameter and uses the
individual format specifications within the FIELD parameter.

WER434I JOINKEYS FORMAT OPERAND IGNORED

EXPLANATION: On the JOINKEYS control statement, the
format of the keys was specified in both the FIELDS and
FORMAT parameters. MFX ignores the FORMAT parameter and
uses the individual format specifications within the FIELDS
parameter.

WER435A ddname ALLOCATION ERROR ON AN UNINITIALIZED

SEQUENTIAL DISK DATA SET
EXPLANATION: An error occurred during the dynamic
allocation of a multi-volume uninitialized sequential disk data
set being used for an input data set. When the UNINTDS=YES
option has been selected, either by default or parameter override,
MFX will need to dynamically allocate and open for output a
multi-volume uninitialized disk data set in order to set the
DS1IND80 flag in the format-1 DSCB of the first volume.
ACTION: In a separate step prior to the MFX invocation, write
the appropriate end-of-file mark in the first volume of the multi-
volume data set.

WER436I UNEQUAL MAINTENANCE APPLIED TO GLOBAL DSM

AND SYNCSORT LIBRARIES
EXPLANATION: The maintenance level of the MFX product is in
conflict with the maintenance level of the global DSM (GDSM)
subcomponent due to the incomplete application of one or more
maintenance levels.

17–50 Syncsort MFX Programmer’s Guide

 WER437A [ddname] SPLIT, SPLITBY, SPLIT1R OR REPEAT
INCOMPATIBLE WITH REPORT WRITING
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. The SPLIT, SPLITBY, SPLIT1R, or REPEAT
parameter and one or more report writing parameters have been
specified for an OUTFIL group. The specified ddname is the first
ddname of the OUTFIL group. SPLIT, SPLITBY, SPLIT1R, or
REPEAT and report writing parameters are incompatible on the
same OUTFIL control statement. Specifically, SPLIT, SPLITBY,
SPLIT1R, or REPEAT cannot be specified on the same OUTFIL
statement with HEADERn, TRAILERn, LINES, NODETAIL,
and SECTIONS.

WER438A [ddname] {INREC,OUTREC} - NONE OF THE FIND-

CONSTANTS WAS MATCHED WITH THE CHANGE FIELD
(p,l), CONTENTS OF INPUT FIELD IN HEX: xxxxxxxx
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. A CHANGE subparameter on an INREC, OUTREC or
OUTFIL OUTREC control statement was specified without a
NOMATCH option and the input field did not match any of the
specified find-constants. p,l represents the position and length of
the input field. xxxxxxxx is the hexadecimal representation of the
input field.

WER439A {INREC, OUTREC, ddname} FIND/REPLACE OVERRUN

OF nnnnn BYTE RECORD LENGTH
EXPLANATION: FINDREP was used on either an INREC,
OUTREC, or OUTFIL ddname statement. The substitution of an
output constant during a FINDREP operation caused a non-
blank character to be pushed beyond the maximum record
length. By default only trailing blanks can be deleted during a
FINDREP operation. nnnnn represents the maximum record
length for the FINDREP operation.
ACTION: Either specify a MAXLEN value on the FINDREP
parameter to increase the maximum length of the record or
specify the OVERRUN=TRUNC FINDREP subparameter to
allow deletion of non-blank characters.

WER440A UNSUPPORTED OPERATING ENVIRONMENT

EXPLANATION: The operating system on which MFX executes
must be z/OS Release 1.4 or later. In addition, MFX requires a
zSeries processor running z/OS in ESAME mode.

Syncsort MFX Programmer’s Guide 17–51

 WER441A ERROR IN CALLING LANGUAGE ENVIRONMENT
SERVICE, RC = nnnn
EXPLANATION: A Language Environment service used to
support LOCALE processing indicated a critical error in its
feedback code. nnnn is the error message number representing
the feedback code. For an explanation of this code, see the IBM
publication Debugging Guide and Run-Time Messages, SC26-
4829.

WER442A INVALID CHARACTER IN COMPARE FIELD FOR

ACTIVE LOCALE
EXPLANATION: INCLUDE/OMIT processing with the LOCALE
function active detected a character that is not defined in the
current locale. The invalid character could be in a CH field or in a
character or hexadecimal constant compared to a CH field.

WER443A INVALID CHARACTER IN CONTROL FIELD FOR

ACTIVE LOCALE
EXPLANATION: Sort or merge processing with the LOCALE
function active detected a character that is not defined in the
current locale. The invalid character is in a CH sort or merge
field.

WER444I LOCALE PROCESSING USED FOR LOCALE nnnnnn

EXPLANATION: Indicates that LOCALE processing was in
effect. nnnnnn (up to 32 characters) represents the name of the
locale used.

WER445A LOCALE PROCESSING CONFLICT

EXPLANATION: LOCALE processing has been used invalidly.
LOCALE processing cannot be used with an E61 exit. The
LOCALE specification cannot be changed on a MAXSORT
breakpoint/restart.

WER446A [ddname] INCLUDE/OMIT FORMATS INCOMPATIBLE

FOR LOCALE PROCESSING
EXPLANATION: The ddname will be SORTOUT, SORTOFxx,
SORTOFx or the ddname provided by an OUTFIL FNAMES
parameter. LOCALE processing has been requested and a
character (CH) to binary (BI) comparison was specified in an
INCLUDE/OMIT or WHEN comparison. CH to BI comparisons
are not supported when using LOCALE processing.

17–52 Syncsort MFX Programmer’s Guide

 WER447B PHASE 3 VIRTUAL STORAGE REDUCED TO nnn BYTES
FOR OPTIMAL PERFORMANCE
EXPLANATION: Phase 3 optimization has determined that a
reduction in virtual storage is appropriate for an efficient
execution. nnn is the amount of virtual storage used during
phase 3. The total bytes used value in message WER164B
indicates the virtual storage used during earlier phases of the
sort execution.

WER448I Y2 FORMAT CENTURY WINDOW IS FROM xxxx TO yyyy

EXPLANATION: One of the Y2x data formats has been used for a
SORT/MERGE field, an INCLUDE/OMIT/WHEN/BEGIN/END
field or an INREC/OUTREC edit field. The starting year is xxxx
and the ending year is yyyy for the century window used to
process the fields.

WER449I SYNCSORT GLOBAL DSM SUBSYSTEM ACTIVE

EXPLANATION: The MFX Global DSM (GDSM) subsystem was
active during the execution of this MFX application.

WER450I PARASORT USED

EXPLANATION: The PARASORT technique has been used for
this execution.

WER451A PARASORT TAPE LABEL ERROR VOL(vvvvvv)

[CONCATENATION+0nnn]
EXPLANATION: The tape label on volume vvvvvv does not
match the DCB characteristics of the input data set. This could
happen because of changed record length, BLKSIZE or record
format. This situation is normally caused by overwriting some of
the data in a multi-volume data set. The concatenation number
indicates where in the input concatenation the volume in error
may be found.

WER452I PARASORT NOT USED: reason

EXPLANATION: The PARASORT feature has been disabled and
the sort was performed using conventional input processing. The
message indicates the reason for this action, which may be any of
the following:
• AUTOMATIC RETRY DISABLED Automatic sort retry
must be enabled for PARASORT to be used. It is required in
the event that the condition identified in WER454A is
encountered.

Syncsort MFX Programmer’s Guide 17–53

 • CONCATENATED SORTIN DEVICES DIFFER
Concatenated SORTIN devices must be the same device; that
is, unit affinity must be specified.
• DUPLICATE VOLUMES ON SORTIN DD NOT ALLOWED
• INCOMPATIBLE CONDITIONS The application may
specify elements that cannot be used together. This problem
can be caused by unusual sort key types, some feature
combinations, or very long sort keys.
• INPUT IS NOT TAPE PARASORT requires input from tape
devices. Input from any other source is not permitted.
• INSUFFICIENT TAPE CHANNELS At least two channel
paths must be available to the tape drives being used to read
the SORTPARn DDs. For a description of a technique to help
insure that this requirement is satisfied, see the description
of esoteric unit names in the PARASORT chapters of this
manual and the MFX for z/OS Installation Guide.
• NO SORTWORKS AVAILABLE PARASORT requires
sortwork space, which must be specified in the JCL or
provided dynamically by DYNALLOC.
• RETRY IN PROGRESS PARASORT failed, but a retry is
being attempted.
• SORTIN IS A NULLFILE
• SORTIN IS ONLY A SINGLE VOLUME DATA SET The
SORTIN DD statement for PARASORT must define either a
single multi-volume SORTIN data set or several concatenated
tape data sets, which can be single or multi-volume. One
single-volume data set is not permitted.
• V(B)S DATA SETS NOT ALLOWED VS and VBS data sets
are not compatible with PARASORT.

WER453A FOR PARASORT text

EXPLANATION: PARASORT failed and the sort application will
not execute. The message text indicates the condition that caused
the failure or the PARASORT requirement that was violated:
• A SORTPAR2 DD STATEMENT IS REQUIRED
• EQUALS MAY NOT BE SPECIFIED If EQUALS is not
specified on the SORT control statement or as a PARM,
ensure it is not enabled by default. Pass NOEQUALS to
disable EQUALS.
• E15 EXITS MAY NOT BE SPECIFIED

17–54 Syncsort MFX Programmer’s Guide

 • IFTHEN WHEN=GROUP MAY NOT BE SPECIFIED ON
INREC
• MAXSORT MAY NOT BE SPECIFIED
• PASSED SORTIN IS INVALID
• SEQNUM MAY NOT BE SPECIFIED ON INREC
• SKIPREC MAY NOT BE SPECIFIED
• SORTIN AND SORTOUT MUST BE DIFFERENT DATA
SETS
• SORTIN GDG NOT ALLOWED
• SORTIN VOLUME SEQUENCE MAY NOT BE
SPECIFIED The volume sequence number must be 1, the
first volume. The number cannot be greater than 1.
• SORTPAR DD STATEMENTS ARE REQUIRED
• SORTPAR(N)S MUST BE SEQUENTIALLY NUMBERED
• SORTPAR1 AND SORTIN DATA SET NAMES MUST BE
THE SAME
• SORTPAR1 DISPOSITION MUST BE OLD
• SORTPAR1 UNIT MUST BE THE SAME AS THE SORTIN
UNIT
• SORTPAR1-4 DEVICE TYPES MUST BE THE SAME AS
THE SORTIN DEVICE TYPE
• SORTPAR2-4 CANNOT BE THE SAME AS THE SORTIN
UNIT
• SORTPAR2-4 and SORTIN DATA SET NAMES MUST BE
THE SAME
• SORTPAR2-4 DISPOSITION MUST BE (NEW,KEEP,KEEP)
• SORTPAR2-4 MUST SPECIFY DEFER ON THE UNIT
PARAMETER
• SORTPAR2-4 MUST SPECIFY VOL=PRIVATE
• STOPAFT MAY NOT BE SPECIFIED
• THE DISPOSITION OF SORTIN IS INVALID SORTIN
data sets may not be temporary data sets. They also may not
be NEW, passed or have PASS on their JCL definition.
• DB2 MAY NOT BE SPECIFIED The DB2 query function is
not supported with a PARASORT.

Syncsort MFX Programmer’s Guide 17–55

 WER454A PARASORT SORTIN END OF FILE ENCOUNTERED
BEFORE THE VOLUME LIST EXHAUSTED
EXPLANATION: The SORTIN volume list is supplied from
either the catalog or specific list of volume serial numbers. The
volume serial list must accurately reflect the volumes in the data
set. If extra volumes are specified (as may happen if an old data
set is rewritten with less data) this error message will be
generated. A volume sequence number may not be specified.

WER455I PARASORT CHANNEL CONTENTION - SORTPARn NOT

USED
EXPLANATION: SORTPAR2-4 has no available channel path to
send data other than a path that would conflict with a previously
defined SORTPARn definition. This SORTPARn will not be used
during the PARASORT execution. This message may occur more
than once if there are multiple conflicting SORTPARn DD’s.

WER456I VISUAL SYNCSORT APPLICATION SUCCESSFULLY

EXPORTED
EXPLANATION: A file that describes your application has been
created and written to the VISUALEX DD statement for export
to Visual SyncSort. The operations defined by the control
statements have not been performed.

WER457A VISUALEX NOT SPECIFIED OR INVALID

EXPLANATION: The VISUALEX DD statement for export to
Visual SyncSort is either missing or its data set has been
incorrectly defined. The file must have physical sequential or
extended sequential organization or be a member of a partitioned
data set or PDSE. The record format must be undefined
(RECFM=U) or unspecified.

WER458A MAINTENANCE LEVEL INSUFFICIENT TO PROCESS

VISUAL SYNCSORT SYSIN DATA SET
EXPLANATION: The SYSIN data set created by Visual SyncSort
cannot be processed by MFX. This is due to an insufficient level of
maintenance on the MFX library. A newer level of MFX may be
required to process the SYSIN data set.

WER459A A VISUAL SYNCSORT APPLICATION MAY NOT text

EXPLANATION: Only qualified MFX applications may be
exported to Visual SyncSort. The reason this application is
ineligible is supplied in the message text.

17–56 Syncsort MFX Programmer’s Guide

 WER460I ddname DATA TRUNCATED DUE TO DCB BLKSIZE
OVERRIDE
EXPLANATION: An extended sequential data set used as input
to a sort, merge or copy has had its DCB BLKSIZE overridden to
a smaller value via a JCL specification. A physical block
exceeding this overridden BLKSIZE specification was truncated
to the smaller size during input processing.
ACTION: Confirm that this truncation is desired. If not, remove
the BLKSIZE specification from the JCL.

WER461A SORTOUT/OUTFIL DATA SET CONTAINS NO DATA

RECORDS
EXPLANATION: If the NULLOUT=RC16 parameter is in effect
and the SORTOUT data set had no data records written to it
during processing, WER461A will be posted. (HEADER/
TRAILER records are not considered to be data records.) If one or
more non-SORTOUT OUTFIL specifications had the
NULLOFL=RC16 parameter in effect and they had no records
written to them, WER461A will be posted. The WER405I
message, which details the records written to each OUTFIL, will
provide information on the OUTFIL(s) that caused the message
to be generated. Note that an OUTFIL FILES=OUT, or FNAMES
SORTOUT is controlled by NULLOUT only, and not by
NULLOFL.

WER461I SORTOUT/OUTFIL DATA SET CONTAINS NO DATA

RECORDS
EXPLANATION: If the NULLOUT=RC4 parameter is in effect
and the SORTOUT data set had no data records written to it
during processing, WER461I will be posted. (HEADER/TRAILER
records are not considered to be data records.) If one or more non-
SORTOUT OUTFIL specifications had the NULLOFL=RC4
parameter in effect and they had no records written to them,
WER461I will be posted. The WER405I message, which details
the records written to each OUTFIL, will provide information on
the OUTFIL(s) that caused the message to be generated. Note
that an OUTFIL FILES=OUT, or FNAMES SORTOUT is
controlled by NULLOUT only, and not by NULLOFL.

WER462A OUTPUT LRECL DIFFERS FROM SORTOUT LRECL

EXPLANATION: If the application is a sort, merge, or copy, the
LRECL defined in the JCL for a non-OUTFIL SORTOUT differs
from the SORTIN/SORTINnn/SORTMInn LRECL or the
internally processed record length when the SORTIN/
SORTINnn/SORTMInn LRECL is modified by features and the
PAD and/or TRUNC parameters have been set to RC=16 to
disallow this. For a variable-length MULTIIN application, the

Syncsort MFX Programmer’s Guide 17–57

 maximum of all of the SORTMInn LRECLs is used. In a
BetterGener application, the LRECL defined in the JCL for
SYSUT2 differs from the SYSUT1 LRECL or the internally
modified record length when the SYSUT1 LRECL is modified by
features and the SOPADGN and/or SOTRNGN installation
options have been set to RC=16 to disallow this.
ACTION: Remove the SORTOUT LRECL specification, allowing
MFX to calculate the appropriate SORTOUT LRECL or modify
the MFX control statements to build a record of the desired
length as specified by the SORTOUT LRECL.

WER462I OUTPUT LRECL DIFFERS FROM SORTOUT LRECL

EXPLANATION: If the application is a sort, merge, or copy, the
LRECL defined in the JCL for a non-OUTFIL SORTOUT differs
from the SORTIN/SORTINnn/SORTMInn LRECL or the
internally processed record length when the SORTIN/
SORTINnn/SORTMInn LRECL is modified by features and the
PAD and/or TRUNC parameters have been set to RC0 or RC4.
For a variable-length MULTIIN application, the maximum of all
of the SORTMInn LRECLs is used. In a BetterGener application,
the LRECL defined in the JCL for SYSUT2 differs from the
SYSUT1 LRECL or the internally modified record length when
the SYSUT1 LRECL is modified by features and the SOPADGN
and/or SOTRNGN installation options have been set to RC=0 or
RC=4.
Fixed-length records will be padded to the SORTOUT LRECL
(SYSUT2 LRECL in a SYNCGENR application) when the
SORTOUT LRECL is greater than the SORTIN or internally pro-
cessed record length.
Records will be truncated to the SORTOUT LRECL (SYSUT2
LRECL in a SYNCGENR application) when the SORTOUT
LRECL is less than the SORTIN or internally processed record
length.
ACTION: Verify that the padding or truncation that will be
performed is desired for this application. Refer to the provided
WER108I and WER110I messages that detail the input and
output record lengths.

WER463A ddname IS A LINEAR VSAM DATA SET

EXPLANATION: MFX does not support an input or output file
that is a linear VSAM data set.

WER464I INVALID SPANNED RECORD FOUND

EXPLANATION: An invalid spanned record segment has been
found while processing the input records in a sort or merge
application, and VLTEST=(,{OFF,OFF4}) has been specified to

17–58 Syncsort MFX Programmer’s Guide

 produce a warning. When OFF4 has been specified, a return code
of 4 will be issued if not overridden by a higher return code issued
for another reason.

WER465A OPEN ERROR SYMNAMES

EXPLANATION: The OPEN for the SYMNAMES DD has failed.
Check the DD statement for any errors.

WER466A SYMNAMES ERRORS FOUND

EXPLANATION: One or more errors were found in the data
dictionary definitions in the SYMNAMES data set or JPn
PARMs. See the description of each error that appears after each
erroneous SYMNAMES statement.

WER467I DB2 QUERY TRIAL MODE SUCCESSFULLY EXECUTED

EXPLANATION: A report of the record layout produced by the
DB2 query contained in the SORTDBIN data set has been
successfully produced. No other processing has occurred.

WER468A DB2 QUERY SUPPORT ERROR: text

EXPLANATION: The DB2 query operation failed and the sort or
copy application will not execute. The message text indicates the
condition that caused the failure or the DB2 query requirement
that was violated.
• MAXSORT MAY NOT BE SPECIFIED
• AN E15 EXIT MAY NOT BE SPECIFIED
• MERGE OPERATION MAY NOT BE SPECIFIED
• SKIPREC MAY NOT BE SPECIFIED
• SORTDBIN OPEN ERROR
• SORTDBIN CANNOT BE FOUND The DB2 parameter has
been specified, but the required SORTDBIN DD has not been
provided.
• NO SQL SELECT STATEMENT FOUND IN SORTDBIN
• INVALID COMMAND, ONLY SQL SELECT
STATEMENT SUPPORTED Only a SELECT or $ELECT
statement is valid in SORTDBIN. No other SQL operations
are supported.
• QUERY STATEMENT TOO LONG (MAX 32765 BYTES)
• CANNOT CONNECT TO DB2 DB2 is not started or the
subsystem name specified on the DB2 EXEC parameter is
incorrect.

Syncsort MFX Programmer’s Guide 17–59

 • CANNOT BIND PLAN The user ID from which the job was
submitted has insufficient authority to bind the plan with the
MFX module. Submit the application from an ID that is
allowed the BIND privilege.
• BIND/OPN PLAN ER The user ID from which the job was
submitted has insufficient authority to bind the plan with the
MFX module or insufficient resources were available for DB2
to process the open request.
• UNSUPPORTED DATA TYPE FOUND
• UNKNOWN DATA TYPE FOUND
• SQL ERROR: SQLCODE=xxxx,SQLSTATE=yyyy Where
xxxx is the SQLCODE and yyyy is the SQLSTATE returned.
Refer to IBM publication DB2 Universal Database for z/OS
Messages and Codes (GC18-9602) for details on these return
codes.
• DB2 MODULES ARE NOT LINKED The DB2 query
facility of MFX has not been installed during MFX
installation. Contact your systems programmer for
assistance.

WER469A BOTH JOINKEYS STATEMENTS MUST HAVE THE SAME

NUMBER OF JOINKEYS
EXPLANATION: A join application requires two JOINKEYS
statements that define the same number of JOINKEYS FIELDS,
with each corresponding field having the same order specified.

WER470A TWO JOINKEYS STATEMENTS ARE REQUIRED TO USE

THE JOIN FEATURE
EXPLANATION: A join application requires two JOINKEYS
statements that define the same number of JOINKEYS FIELDS,
with each corresponding field having the same order specified.

WER471A A REFORMAT STATEMENT IS REQUIRED TO USE THE

JOIN FEATURE
EXPLANATION: A join application requires two JOINKEYS
control statements and a REFORMAT control statement, unless
a JOIN UNPAIRED control statement is present with the ONLY
parameter specified.

WER472A THE NUMBER OF JOIN FIELDS EXCEEDS THE

MAXIMUM OF 64
EXPLANATION: The maximum number of JOINKEYS FIELDS
that may be specified is 64.

17–60 Syncsort MFX Programmer’s Guide

 WER473A INVALID JOINKEYS STATEMENT FIELD LENGTH
EXPLANATION: A JOINKEYS statement field length exceeds
the allowable length for its format. 4080 is the maximum for BI,
CH and AQ fields, and 256 is the maximum for FI, PD, and ZD
fields.

WER474A TOTAL LENGTH OF JOIN KEYS IN JOINKEYS

STATEMENTS EXCEEDS 4080
EXPLANATION: The maximum total length of all JOINKEYS
FIELDS is 4080 bytes.

WER475A BOTH JOINKEYS STATEMENTS MUST HAVE

CORRESPONDING KEY FIELDS WITH COMPATIBLE
FORMATS
EXPLANATION: Corresponding JOINKEYS fields do not have
formats that are compatible with each other. CH and BI are
compatible formats, and PD and ZD are compatible formats. AQ
and FI formats are not compatible with any other format type.

WER476A FIRST 4 BYTES OF A VARIABLE-LENGTH REFORMAT

DEFINITION MUST BE FROM A VL JOIN FILE
EXPLANATION: When defining a variable-length record as the
output of the join feature, the first four bytes defined must be for
an RDW. This must be specified on the REFORMAT statement as
FIELDS=(Fn:1,4,…). This field may be taken from either of the
SORTJNFn files if both are variable-length, but if only one file is
variable-length, the field must come from the variable-length file.

WER477A BOTH JOINKEYS STATEMENTS MUST HAVE

CORRESPONDING KEY FIELDS WITH THE SAME
ORDER
EXPLANATION: A join application requires two JOINKEYS
statements that define the same number of JOINKEYS FIELDS,
with each corresponding field having the same order specified.

WER478A A SORTJNF1 OR SORTJNF2 DD STATEMENT IS

MISSING; BOTH ARE REQUIRED
EXPLANATION: A join application requires the presence of
SORTJNF1 and SORTJNF2 DD definitions in the JCL.

WER479A xxxxxxxx MAY NOT BE USED IN A JOIN APPLICATION

EXPLANATION: A join application may not be specified with any
of the following parameters: MAXSORT, PARASORT, Syncsort
PipeSort, SKIPREC, MERGE function, user exits (except E35),
CHECKPOINT, or DB2.

Syncsort MFX Programmer’s Guide 17–61

 WER480A A REFORMAT FIELD WITH ONLY A POSITION VALUE
MUST REFERENCE A VARIABLE-LENGTH JOIN FILE
EXPLANATION: A REFORMAT FIELD may only be specified as
a position without a length value if it refers to a file defined as
variable-length. This type of field definition may be specified once
for each join file if they are both variable-length. These
specifications must be the last fields defined on a REFORMAT
statement.

WER481I JOINKEYS REFORMAT RECORD LENGTH = nnnnn,

TYPE = {F,V}
EXPLANATION: nnnnn represents the length of the record
produced by JOINKEYS REFORMAT processing. The TYPE
represents the record format produced, either fixed (F) or
variable (V). If you have variable-length records, nnnnn
represents the maximum record length.

WER482I JNFn STATISTICS

EXPLANATION: JNFn is either JNF1 or JNF2 representing
information on SORTJNF1 or SORTJNF2 processing that was
performed during a join application. This message will be
followed by other informational messages that apply to the SORT
or COPY processing for that DD.

WER483B JNF1 or JNF2 processing information …

EXPLANATION: This message follows the WER482I message
that defines which of the two join input files this message
contains information about. The information details
characteristics of the sort processing performed for SORTJNF1 or
SORTJNF2 to prepare them for the join operation. Further
information can be found in the explanations of the WER164B,
WER410B, WER036B, WER158I, WER460I, WER464I, and
WER162B messages.

WER484I ddname: RCD IN=aaaaaaaa, OMITTED=bbbbbbbb,

PAIRED=cccccccc, UNPAIRED=dddddddd
EXPLANATION: The ddname will be either SORTJNF1 or
SORTJNF2. aaaaaaaa represents the number of records read
from the particular join input file. bbbbbbbb indicates the
number of records deleted by the JOINKEYS INCLUDE/OMIT
parameter. cccccccc is the number of records matched from this
file during join processing. dddddddd is the number of records
from this file that were unpaired during join processing.

17–62 Syncsort MFX Programmer’s Guide

 WER485A SORTJNFn OUT OF SEQUENCE, RECORD
NUMBER=aaaaaaaa
EXPLANATION: When the SORTED parameter is specified on
the JOINKEYS statement, MFX will sequence check the input
file according to its JOINKEYS fields. This message indicates
that a sequence error was detected in either SORTJNF1 or
SORTJNF2. aaaaaaaa is the record number within the file that
caused the error. The record number is not provided if the
INCLUDE/OMIT parameter of the JOINKEYS statement is
specified.

WER486A ERROR IN JNFn PROCESSING

EXPLANATION: The n value is either 1 or 2 and indicates that
an error has occurred while processing the SORTJNF1 or
SORTJNF2 data sets. Examine the application’s SYSOUT
message data set for additional WERnnnA messages that will
indicate the exact nature of the error.

WER487I FILESIZE aaaaaaaa BYTES

EXPLANATION: The aaaaaaaa represents the number of bytes
of SORTJNF1 or SORTJNF2 input data sorted by MFX to
prepare the file for join match field processing. This number
reflects deletions made by JOINKEYS INCLUDE/OMIT
processing. See the prior WER482I message to determine if
FILESIZE is for SORTJNF1 or SORTJNF2.

WER488A JOIN CAPACITY EXCEEDED

EXPLANATION: MFX is unable to complete the join application
because of insufficient memory. This may be due to a user error in
specifying the JOINKEYS fields, or the application may be very
large relative to the amount of available memory.
In order to join all records with equal JOINKEYS in SORTJNF1
with all records with matching JOINKEYS in SORTJNF2, MFX
retains all the equally keyed SORTJNF2 records in memory to
join with the next equally keyed record from SORTJNF1. The
available amount of memory is determined by the available sys-
tem resources and the region size and may not be sufficient if
there are very many equally keyed records in SORTJNF2.
ACTION: First examine the fields specified for the JOINKEYS
FILE=F2 statement and correct any errors. If the fields were
incorrectly specified, then many records may have been
incorrectly determined to be equally keyed.
If the JOINKEYS statement for F2 is correct, and if you believe
that SORTJNF1 does not contain many equally keyed records
that match the large number of equally keyed records in SORT-
JNF2, then this problem may be easily corrected by reversing the

Syncsort MFX Programmer’s Guide 17–63

 F1 and F2 definitions such that SORTJNF1 has many equally
keyed records, but SORTJNF2 does not. To do this, reverse the
DDNAMEs of the two files and change the JOINKEYS, REFOR-
MAT, and JOIN UNPAIRED statements accordingly.
If this does not solve the problem, then the total region size must
be raised high enough to contain all of the SORTJNF2 equally
keyed records.
Contact Precisely Support for assistance, if necessary.

WER489A RECORD STATEMENT LENGTH VALUE EXCEEDS

MAXIMUM OF 32767
EXPLANATION: The maximum value that can be specified for
an l1 to l7 value is 32767.

WER490I INVALID DATE ENCOUNTERED IN DATE FORMAT

ARITHMETIC OR CONVERSION
EXPLANATION: An INREC/OUTREC field with a Y2x or Y4x
full date format or a DATEADD/DATEDIFF date field contains
an invalid date. The output will be presented with all 9’s in the
digit portion of the specification.

WER491A Z/ARCHITECTURE ENVIRONMENT REQUIRED

EXPLANATION: FL format conversion on INREC, OUTREC, or
OUTFIL is only supported in z/Architecture mode on zSeries
processors.

WER492A DUPKEYS: text

EXPLANATION: A DUPKEYS control statement was specified.
The message text indicates the reason for the error message.
• AVG RECORD COUNT OVERFLOW MFX’s default
internal limit on the maximum number of records that can be
averaged has been exceeded, or the internal accumulation
limit on the summation of the total has been exceeded. By
default, the internal limit on the number of records that can
be processed for an average application is 2,147,483,647
records containing the same SORT or MERGE control fields.
• INVALID xxx DATA FIELD where xxx=SUM, MAX, MIN,
or AVG. A field with an invalid data length was specified in
the xxx parameter.
• INVALID OVERLAPPING OF xxx FIELDS where
xxx=SUM, MAX, MIN, or AVG. An xxx field overlaps another
xxx field or other field in the DUPKEYS statement, or a
SORT/MERGE control field, or the Record Descriptor Word of
a variable-length record. All of these are invalid.

17–64 Syncsort MFX Programmer’s Guide

 • INVALID USE FOR SUM/XSUM The SUM control
statement and the DUPKEYS control statement cannot both
be specified. If you would like to add the MIN, MAX, or AVG
functionality to an application with a SUM control statement,
then move the SUM specification to the DUPKEYS statement
and remove the SUM statement. If XSUM was used, then
XDUP should be specified and the JCL changed from using a
SORTXSUM DD to a SORTXDUP DD.
• INVALID USE FOR XDUP/FORMAT Another parameter
in addition to XDUP or FORMAT must be specified on the
DUPKEYS statement.
• SUM/AVG FIELD OVERFLOW Summing or averaging of
two equally keyed records could not be done due to a numeric
overflow or underflow in a defined SUM/AVG field. The
WER492A critical message is issued instead of the WER492I
warning message if the OVFLO=RC16 PARM or installation
SUMOVFL=RC16 is in effect.
• SUM/MIN/MAX/AVG FIELD OUTSIDE RANGE A field
in the SUM/MIN/MAX/AVG parameter is located beyond the
record length.

WER492I DUPKEYS: text

EXPLANATION: A DUPKEYS control statement was specified.
The message text indicates the reason for the informational
message.
• CONTROL STATEMENT IGNORED The DUPKEYS
control statement is ignored in a FIELDS=COPY or
BetterGener application because there are no SORT/MERGE
control fields.
• FORMAT OPERAND IGNORED The FORMAT
parameter of DUPKEYS was ignored because all fields had a
format specified.
• SHORT RECORD FOR SUM/AVG/MIN/MAX One or
more variable-length records were too short to contain all the
control fields specified on the DUPKEYS statement. No
DUPKEYS function will be performed on this record. The
HISTOGRM program may be used to determine the length of
the shortest record in the input file.
• SUM/AVG FIELD OVERFLOW Summing or averaging of
two equally keyed records could not be done due to a numeric
overflow or underflow in a defined SUM/AVG field. The
WER492A critical message is issued instead of the WER492I

Syncsort MFX Programmer’s Guide 17–65

 warning message if the OVFLO=RC16 PARM or installation
SUMOVFL=RC16 is in effect.

WER493I ZIIP PROCESSOR USED

EXPLANATION: MFX has used the zIIP processor for improved
performance.

WER494I {INPUT, OUTPUT} PHASE USED MIDAW {;MIXED MODE}

EXPLANATION: MFX’s MIDAW technology optimized the
performance of the input and/or output phase. MIXED MODE
indicates that some of the devices used are not MIDAW-capable.

WER495A SORTOUT/OUTFIL DATA SET CONTAINS DATA

RECORDS
EXPLANATION: If the NOTMTOUT=RC16 parameter is in
effect and the SORTOUT data set had at least one record written
to it during processing, WER461A will be posted. If one or more
non-SORTOUT OUTFIL specifications had the
NOTMTOFL=RC16 parameter in effect and they had at least one
record written to them, the WER495A will be posted. The
WER405I message, which details the records written to each
OUTFIL, will provide information on the OUTFIL(s) that caused
the message to be generated. Note that an OUTFIL FILES=OUT,
or FNAMES SORTOUT is controlled by NOTMTOUT only, and
not by NOTMTOFL.

WER495I SORTOUT/OUTFIL DATA SET CONTAINS DATA

RECORDS
EXPLANATION: If the NOTMTOUT=RC4 parameter is in effect
and the SORTOUT data set had at least one record written to it
during processing, WER495I will be posted. If one or more non-
SORTOUT OUTFIL specifications had the NOTMTOFL=RC4
parameter in effect and they had at least one record written to
them, WER495I will be posted. The WER405I message, which
details the records written to each OUTFIL, will provide
information on the OUTFIL(s) that caused the message to be
generated. Note that an OUTFIL FILES=OUT, or FNAMES
SORTOUT is controlled by NOTMTOUT only, and not by
NOTMTOFL.

WER496A THE MULTIIN PARAMETER CANNOT BE USED WITH

MERGE, PARASORT OR JOIN
EXPLANATION: MERGE, PARASORT and JOIN do not support
the multiple input feature. Specify SORTINnn DD statements for
a merge application.

17–66 Syncsort MFX Programmer’s Guide

 WER497A statement: {NOT-A-NUMBER, NOT-A-NUMBER OR
INFINITY} ENCOUNTERED IN FD FIELD (nnnn,nnnn)
EXPLANATION: For a SORT/MERGE statement, a NOT-A-
NUMBER value is detected in a field (nnnn,nnnn); for a
DUPKEYS/SUM statement, either a NOT-A-NUMBER or
INFINITY value is detected in a field (nnnn,nnnn).

WER498A DECIMAL FLOATING POINT FACILITY REQUIRED FOR

FD FIELDS
EXPLANATION: For a statement that includes a decimal
floating point (DFP) field (FD), the DFP facility is required to be
installed in the z/Architecture architectural mode.

WER499A PFPO SUPPORT REQUIRED TO CONVERT FROM FL TO

FD
EXPLANATION: For an INREC/OUTREC statement that needs
to convert an FL field to an FD field, the PFPO instruction is
required to be installed.

WER500I SYNCSORT STATISTICS DATA SET NOW OVER xx

PERCENT FULL
EXPLANATION: xx percent of space currently allocated on the
MFX Statistics data set has been used. This message is not
controlled by the MSG or FLAG PARM and will appear only on
the console.

WER501A SYNCSORT STATISTICS DATA SET NOW FULL - NO

RECORD WRITTEN
EXPLANATION: The MFX Statistics data set did not have
enough space for the SYNCSMF record. This message is not
controlled by the MSG or FLAG PARM and will appear only on
the console.

WER502A COULD NOT FIND OR LOAD ONE OR MORE JAVA

CLASSES
EXPLANATION: One or more JAVA classes cannot be found.
First, ensure that both the JZOS and MFX libraries have been
installed correctly. Then ensure that JZOSHOME and
SYNCHOME point to these libraries correctly.

WER503A ddname OUTFIL OUTPUT PARAMETER - JAVA

EXCEPTIONS DETECTED
EXPLANATION: JAVA exceptions have been detected. See the
messages in the STDERR DD or the STDOUT DD for the
exception type.

Syncsort MFX Programmer’s Guide 17–67

 WER504A UNABLE TO DYNALLOC WORK DATA SET FOR OUTFIL
PROCESSING
EXPLANATION: Dynamic allocation of the work data set
STDENV failed.
ACTION: Contact Precisely Support or call your MFX support
representative.

WER506I PDF/RTF OUTPUT LINES WRAPPED

EXPLANATION: Some output lines are longer than the page
width. They are wrapped to the following line.

WER507A PDF/RTF/HTML OUTPUT HFS FILES REQUIRED

EXPLANATION: PDF/RTF/HTML output must be HFS.

WER508A OUTFIL JAVA FILE SSOUTPUT.JAR LEVEL DOES NOT

MATCH SYNCSORT
EXPLANATION: The JAR file ssoutput.jar which is installed on
UNIX is not the same level as other MFX modules on the
mainframe. Re-install ssoutput.jar or check if SYNCHOME
points to the correct file directory.

WER509A SORTMInn CANNOT BE CONCATENATED INPUT

EXPLANATION: SORTMInn is a concatenated input data set;
use a SORTMInn DD statement for each of the concatenated
files.

WER510A OUTFIL OUTPUT PARAMETER PROCESSING

CONFLICT
EXPLANATION: LOCALE, C exits and COBOL exits are not
supported with the OUTFIL OUTPUT feature.

WER511A OUTFIL OUTPUT PARAMETER JAVA ENVIRONMENT

PROBLEM
EXPLANATION: An unexpected return from the JAVA
environment occurred. This can occur if the CPU time limit has
expired. Increase the amount of time in the TIME parameter.

WER513A ERRORS IN SYMNAMES STATEMENTS

EXPLANATION: Errors have been found in one or more
SYMNAMES statements or JPn PARMs. See additional
messages for details.

WER514A MISSING MATCHING QUOTE

EXPLANATION: The data dictionary statement or JPn PARM
contains an open quote that has no matching closing quote.

17–68 Syncsort MFX Programmer’s Guide

 WER515A SYMBOL IS A RESERVED WORD
EXPLANATION: The symbol to be defined is a reserved word.

WER516A DUPLICATE SYMBOL DEFINITION

EXPLANATION: The symbol to be defined has been defined
previously.

WER517A UNKNOWN SYMBOL REFERENCED

EXPLANATION: The POSITION data dictionary statement
refers to an undefined symbol.

WER518A POSITION/LENGTH VALUE OUT OF RANGE

EXPLANATION: The data dictionary statement contains a
position or length field that has an invalid value.

WER519A SYMBOL/CONSTANT TOO LONG

EXPLANATION: The symbol is too long or the length of the
constant is invalid for that type of constant.

WER520A INVALID CHARACTER IN CONSTANT DEFINITION

EXPLANATION: The constant contains a character that is
invalid for that type of constant.

WER521A INVALID FORMAT CODE

EXPLANATION: The format specified is not one of the valid data
formats.

WER522A SYNTAX ERROR

EXPLANATION: The data dictionary statement has a syntax
error.

WER523A OUTFIL OUTPUT PARAMETER REQUIRES SYNCSORT

INSTALLATION OPTIONS FOR JAVA ENVIRONMENT
EXPLANATION: The JAVAHOME, JZOSHOME and
SYNCHOME installation options have to point to the correct
libraries when the OUTFIL OUTPUT feature is used.

WER524A MODULE JVMLDM NOT FOUND, JAVA JZOS

ENVIRONMENT NOT AVAILABLE
EXPLANATION: Either JZOS LOADLIB was not installed or it
was not defined in the JOBLIB/STEPLIB data set.

WER525A UNABLE TO SEND EMAIL, SEE JAVA MESSAGES

EXPLANATION: Email cannot be sent. See the messages in the
STDERR DD or the STDOUT DD for the reason.

Syncsort MFX Programmer’s Guide 17–69

 WER526A CANNOT BRING UP JAVA ENVIRONMENT WITHOUT
SYNCSORT SVC IN EFFECT
EXPLANATION: The MFX SVC number must be defined in your
installation default options or passed as a runtime parameter. In
addition the SVC module must be in LPA.

WER527A INVALID CONTROL STATEMENT FOR JOIN INPUT

PROCESSING
EXPLANATION: A control statement in the xxxxCNTL DD file is
not permitted in a join application. The JOINKEYS, JOIN,
MERGE, OUTFIL, OUTREC, REFORMAT and SORT
statements are prohibited because they are inappropriate for the
subtask that reads a JOINKEYS input file.

WER528A IFTRAIL TRLUPD INVALID COLUMN

EXPLANATION: The TRLUPD subparameter of the OUTFIL
IFTRAIL parameter has specified a field in columns 1 through 4
of a variable-length record. This is not permitted because the
field would overlay the RDW of the record. The first defined field
must start in column 5 or beyond.

WER529I IOCB: nnn IOXB: nnn SIOB: nnn CC: nn CSW: nnn SNS:
nnn RBC:nnn DCWO:nnn PCRC:nnn PCRQ:nnn
LREA:nnnnnnnnn
EXPLANATION: This message is issued as a WTO message by a
zHPF subtask upon an unrecoverable I/O error. It contains
internal diagnostic information that is intended for use by
Precisely Support personnel.

WER530I UNRECOVERABLE TRANSPORT-MODE (ZHPF) ERROR

ON DEVICE xxxx; REVERTING TO COMMAND-MODE
FOR yyyyyyy
EXPLANATION: This message is issued when an unrecoverable
zHPF error occurred and I/O processing reverted back to
command-mode. xxxx is the device number on which the error
occurred, and yyyyyyyy is the DD Name for either SORTIN or
SORTOUT.

WER559I UNSUPPORTED FUNCTIONS WITH SORTWORK

ENCRYPTION
EXPLANATION: SORT is not using encryption mode for
SORTWORK because of unsupported functions.

WER560I SORTWORK ENCRYPTION USED

EXPLANATION: SORT is using encryption mode for
SORTWORK. All data in the SORTWORK data set is encrypted.

17–70 Syncsort MFX Programmer’s Guide

 WER561A IBM CALLABLE SERVICE XXXXXXXX ERROR, RETURN
CODE XXXXXXXX, REASON CODE XXXXXXXX
EXPLANATION: A SORTWORK encryption process error
occurred during an IBM call statement.
ACTION: For more information about the IBM return and reason
codes, refer to the IBM publication z/OS Cryptographic Services
Integrated Cryptographic Service Facility Application
Programmer’s Guide, SC14-7508-06. Or contact Precisely
Support for assistance.

Syncsort ZPCopy Messages

WER550I ZPCOPY EXECUTED - TYPICAL SAVINGS ARE UP TO
95% TCB TIME AND 20% ELAPSED TIME
EXPLANATION: This message appears when ZPCopy is in
effect.

WER551I ZPSAVER IS ATTEMPTING TO RUN IN AN

UNAUTHORIZED ENVIRONMENT. ZPSAVER IS
DISABLED.
EXPLANATION: Syncsort ZPSaver could not be used because
the library or module is not APF-authorized.
ACTION: Determine why the library or module is not authorized
and correct the problem so Syncsort ZPSaver will be enabled. The
most frequent causes are:
• Syncsort ZPSaver is executing out of an unauthorized
library.
• Certain Syncsort MFX load modules were not relinked
with the load module attribute AC=1.
• An unauthorized library was included in the JOBLIB or
STEPLIB concatenation.

Syncsort ZPCompress Messages

WER553I ZPCOMPRESS USED FOR SORTOUT
EXPLANATION: This message indicates that ZPCompress was
used for SORTOUT compression.

Syncsort MFX Programmer’s Guide 17–71

 WER554I ZPCOMPRESS USED FOR SORTIN
EXPLANATION: This message indicates that ZPCompress is
used for SORTIN decompression.

Syncsort ZPSort Messages

WER558I ZPSORT EXECUTED - x% CPU TIME SAVED,
ADDITIONAL y% COULD HAVE BEEN SAVED IF ZIIP
ENGINES WERE NOT BUSY
EXPLANATION: Syncsort ZPSaver was executed. x% of CPU
time has been saved as a result of zIIP offload. Because the zIIP
engines were busy, y% of CPU time that could have been
offloaded to zIIP was offloaded to a regular CP.

Syncsort PROCSort Messages

WER700A PROC SYNCSORT UNSUPPORTED FUNCTION.
{RETRY,NORETRY} IN EFFECT
EXPLANATION: MFX’s high performance technique could not be
used during this invocation by Syncsort PROCSort - An
Accelerator for SAS Sorting. This may be due to a small region
size or the generation of an unsupported MFX statement syntax.
If the RETRY option of Syncsort PROCSort is in effect, MFX will
be reinvoked using a less efficient E15-E35 interface. If the
RETRY option is not in effect, the Syncsort PROCSort execution
will be terminated.

WER744A CONFLICT BETWEEN SYNCSORT AND PROC

SYNCSORT MAINTENANCE LEVELS, VERIFY
LIBRARIES
EXPLANATION: Maintenance has been applied to either
Syncsort PROCSort or MFX, but not to both when maintenance
to both is required.
ACTION: Check the libraries containing Syncsort PROCSort and
MFX and apply the required level of maintenance to each.

WER775A SAS I/O ERROR OCCURRED. CHECK SAS MESSAGE LOG

DATA SET
EXPLANATION: An I/O error occurred when a SAS routine
attempted to access or update a SAS data set. A message

17–72 Syncsort MFX Programmer’s Guide

 indicating the actual nature of the problem should appear on the
SAS message LOG data set.

WER776A BLDL FAILURE FOR DDNAME SASLIB. RAISE REGION

OR CHECK SASLIB ACCESS
EXPLANATION: When attempting to perform a BLDL for the
library identified by the SASLIB DD statement, an error
occurred. The error is due either to insufficient virtual storage or
a permanent I/O error on the library.

WER777A ERROR LOADING PROC SYNCSORT MODULE. CHECK

PROC SYNCSORT INSTALL
EXPLANATION: The Syncsort PROCSort module could not be
found in any of the libraries on the normal z/OS search chain or
an error occurred while loading the module.

WER778A UNEQUAL MAINTENANCE APPLIED TO PROC

SYNCSORT AND SYNCSORT LIBRARIES. DATA=hexdata
EXPLANATION: The maintenance level of the Syncsort
PROCSort product is in conflict with that of the MFX product due
to the incomplete application of one or more maintenance levels.
The hexadecimal data, if printed, indicates which maintenance
fixes were incompletely applied.

WER779I THE PERFORMANCE OF THIS SORT COULD BE

SIGNIFICANTLY IMPROVED THROUGH THE USE OF
THE PROC SYNCSORT PRODUCT
EXPLANATION: Syncsort PROCSort - An Accelerator for SAS
Sorting is a high performance replacement for the SAS-provided
procedure PROC SORT. When MFX is invoked with the Syncsort
PROCSort product instead of through the interface supplied by
SAS, significant performance improvements result. For more
information, call Precisely Support.

License Key Messages

The following are the messages directly related to the use of license keys for the
Syncsort MFX, Syncsort PROCSort, and Syncsort PipeSort products.

WER900A SYNCSORT r.r.n.n IS NOT LICENSED FOR SERIAL sssss,

TYPE mmmm mmm, [LPAR nn,] MSU ccccc.
or

Syncsort MFX Programmer’s Guide 17–73

 SYNCSORT r.r.n.n IS NOT LICENSED FOR SERIAL sssss,
TYPE mmmm, [LPAR nn,] VERSION CODE vv.
EXPLANATION: r.r is the MFX release number, and n.n is the
TPF maintenance level. A different product name is displayed
when applicable. No valid license key for use on the specified
machine was found, and the grace period for this error, noted by
the WER903I warning message, has expired. A key must contain
the correct information for both the serial number and the
machine capacity. License keys are specified either in the KEY
parameter of the SYNCMAC installation options macro, or
included in a data set whose name is specified in the KEYDSN
parameter of SYNCMAC.
ACTION: Execute the SYNCLIST program on the system where
this message is occurring. Ensure that either the SYNCMAC
KEY parameter or the data set named in the KEYDSN
parameter has provided a valid key for the specified product for
this machine. If you require further assistance, contact Precisely
Support with the SYNCLIST output available for reference.

WER901I **WARNING** SYNCSORT r.r.n.n WILL EXPIRE IN nnn

DAYS
EXPLANATION: r.r is the MFX release number, and n.n is the
TPF maintenance level. A different product name is displayed
when applicable. The provided license key for this machine is
only valid for the next nnn days. After that time, WER902A will
be issued, and the specified product cannot be used.
ACTION: Contact the systems programmer in charge of MFX
maintenance, or execute the SYNCLIST program on the system
where this message is occurring and contact Precisely Support.

WER902A SYNCSORT r.r.n.n HAS EXPIRED

EXPLANATION: r.r is the MFX release number, and n.n is the
TPF maintenance level. A different product name is displayed
when applicable. The provided license key for this machine is no
longer valid because the expiration date has passed. The
specified product can no longer be used.
ACTION: Contact the systems programmer in charge of MFX
maintenance, or execute the SYNCLIST program on the system
where this message is occurring and contact Precisely Support.

17–74 Syncsort MFX Programmer’s Guide

 WER903I SYNCSORT r.r.n.n IS NOT LICENSED FOR SERIAL sssss,
TYPE mmmm mmm, [LPAR nn,] MSU ccccc.

~or~

SYNCSORT r.r.n.n IS NOT LICENSED FOR SERIAL sssss,

TYPE mmmm, [LPAR nn,] VERSION CODE vv.

SYNCSORT WILL STOP WORKING IN nnn DAYS UNLESS

A VALID KEY IS INSTALLED.
EXPLANATION: r.r is the MFX release number, and n.n is the
TPF maintenance level. A different product name is displayed
when applicable. No valid license key for use on the specified
machine was found. License keys are specified in the KEY
parameter of the SYNCMAC installation options macro, or
included in a data set whose name is specified in the KEYDSN
parameter of SYNCMAC.
Processing continues by issuing WER903I during a grace period
after this error is first encountered. This will provide sufficient
time to correct the problem by installing a valid key for this
machine. If the grace period ends before a valid key is made
available, either WER900A or WER902A will be issued and pro-
cessing will terminate.
ACTION: Execute the SYNCLIST program on the system where
this message is occurring. Ensure that either the SYNCMAC
KEY parameter or the data set named in the KEYDSN
parameter has provided a valid key for this machine. If you
require further assistance, contact Precisely Support with the
SYNCLIST output available for reference.

WER904I SYNCSORT r.r.n.n KEYUPDATE SUCCESSFUL;

xxxxxxxxxxxxxxxx SELECTED
EXPLANATION: r.r is the MFX release number, and n.n is the
TPF maintenance level. A different product name is displayed
when applicable. The KEYUPDATE parameter was specified,
and MFX has successfully obtained a valid license key denoted by
xxxxxxxxxxxxxxxx from MFX’s key data set. The name of the
data set was specified in the KEYDSN parameter of the
SYNCMAC installation options macro.

WER905A SYNCSORT r.r.n.n KEYUPDATE FAILURE: reason

EXPLANATION: r.r is the MFX release number, and n.n is the
TPF maintenance level. A different product name is displayed
when applicable. The KEYUPDATE parameter was specified, but
MFX was unable to obtain a valid license key from MFX’s key
data set due to the specified reason. Possible reasons for this
failure are:

Syncsort MFX Programmer’s Guide 17–75

 1. The KEYDSN parameter of SYNCMAC was not specified
when MFX was installed. KEYDSN, and not the KEY
parameter, must be specified with the name of MFX’s key
data set when using the KEYUPDATE facility.
2. MFX was unable to dynamically allocate and/or read MFX's
key data set. This can happen if you were editing the data set
at the time of the KEYUPDATE run, or if the data set was not
allocated as a fixed-length 80-byte file.
3. No valid license key was found in MFX’s key data set.
4. The MFX SVC was not available. MFX requires use of its SVC
to perform the update.
ACTION: Ensure that the KEYDSN parameter has been
correctly specified and that the data set is accessible and contains
a valid license key. Also verify that the MFX SVC has been
properly installed. If you require further assistance, execute the
SYNCLIST program on the system where this message is
occurring and contact Precisely Support with the SYNCLIST
output available for reference.

WER906I INVALID KEY DATA SET RECORD:

invalid record text

EXPLANATION: One or more invalid records were found in the
license key data set when performing KEYUPDATE. The first
invalid record is displayed in the message text. Only comment
statements, key statements and valid PARMS statements are
permitted. All invalid statements are ignored.
ACTION: Correct any errors in the key data set record that was
displayed in the message text and rerun the KEYUPDATE
application.

WER907I SYNCSORT EXPIRING LICENSE KEY WARNING

MESSAGE {ENABLED,DISABLED}

~or~

SYNCSORT INVALID LICENSE KEY WARNING

MESSAGE {ENABLED,DISABLED}
EXPLANATION: These KEYUPDATE messages document
whether MFX may issue certain license key warning messages.
These messages also apply to Syncsort PROCSort and Syncsort
PipeSort if these products have been installed. The default is to
issue either the WER901I expiring license key warning message
or the WER903I invalid license key warning message when
applicable. During KEYUPDATE, a PARMS statement read from
the key data set can disable the issuance of either of these

17–76 Syncsort MFX Programmer’s Guide

 messages. The WER907I message is intended to alert you that
these warning messages may no longer be posted, though the
warning period countdowns will continue. During the last seven
days before the warning period ends, the warning messages are
issued regardless of whether or not they have been disabled.
This is done to try to prevent termination of all applications with
either WER902A or WER900A.
ACTION: No action is required if both of these warning messages
are enabled and you have a valid license key that is not expiring.
If you do not have a valid key or if your key is expiring, call
Precisely Support as soon as possible to obtain a new license key
and rerun the KEYUPDATE procedure using the new key. If any
of the messages had been disabled, either remove the PARMS
statement or set the warning message parameters to ON to re-
enable the issuance of license key warning messages.

WER908A DEGRADED PROC SYNCSORT PERFORMANCE! CALL

YOUR SYNCSORT REPRESENTATIVE.
EXPLANATION: WER900A or WER902A has been issued
because there is no valid license key for Syncsort PROCSort.
Processing will continue, but MFX’s high performance technique
will not be used during this invocation by Syncsort PROCSort.
ACTION: See EXPLANATION and ACTION for either WER900A
or WER902A, as appropriate.

Syncsort MFX Programmer’s Guide 17–77

17–78 Syncsort MFX Programmer’s Guide
Chapter 18 Diagnostics and Technical
Support

Troubleshooting Abends
Troubleshooting with WER999A UNSUCCESSFUL SORT
WER999A indicates that an error condition occurred, preventing the successful
completion of the sort. This message does not necessarily mean that MFX was
responsible for the error. If, for example, the error is in the COBOL Input or Output
Procedure of an invoked sort, WER999A will appear. WER999A indicates that MFX
got control after the error, printing this MFX message.
The documentation accompanying WER999A varies with the error involved. It may
consist of a standard system dump (SYSUDUMP or SYSABEND) and/or an MFX-
generated SNAP dump. The MFX SNAP is formatted very much like a
SYSUDUMP. In debugging the SNAP, care must be taken to avoid reliance on the
PSW AT ENTRY TO SNAP and the general registers.
A SNAP dump produced through the MFX DEBUG PARM or with a W-abend (that
is, WER999A UNSUCCESSFUL SORT xxxW) is only useful to a sort analyst at
Precisely Support. See “Before Contacting Precisely Support” later in this chapter.

MFX Internal Abend

A W-type abend code indicates that program termination was forced by an error
condition internally detected by MFX; the problem cannot be resolved by the user.
See “Before Contacting Precisely Support” later in this chapter.

U-Type Abend Codes

If any of the U-type abend codes in the chart below appears in the WER999A
message, it may indicate an MFX error (in which case, see “Before Contacting
Precisely Support”). These are the only U-type abend codes that MFX issues; any

Syncsort MFX Programmer’s Guide 18–1

 U-type abend code which is not on this list indicates an error in a user-written exit
routine, invoking program or environment. For example, user abend 4093 (RC=1C)
is related to LOCALE processing. This abend is issued from the LE/370
environment when the REGION is not large enough. To address a U4093 abend,
increase the REGION by 1 megabyte and resubmit the application.

Note that the WER999A message displays the abend code in hexadecimal.

User-Type Abend Codes Issued by MFX

_______________________________________
Decimal Hexadecimal
10 A
16 * 10 *
54 36
69 45
100 64
101 65
200 C8
300 12C
400 190
936 3A8
999 ** 3E7 **
1024 *** 400 ***
1025 401
1026 402
1027 403
1028 404
1030 406
1031 407
1032 408
1050 41A
1051 41B
1052 41C
1053 41D
1054 41E

Table 69. (Page 1 of 3) User Type Abend Codes

18–2 Syncsort MFX Programmer’s Guide

 User-Type Abend Codes Issued by MFX
_______________________________________
Decimal Hexadecimal
1060 424
1061 425
1070 42E
1071 42F
1072 430
1073 431
1074 432
1075 433
1076 434
1077 435
1078 436
1079 437
1086 43E
1087 43F
1088 440
1102 44E
1103 44F
1104 450
1107 453
1108 454
1110 456
1111 457
1112 458
1115 45B
1116 45C
1117 45D
1118 45E
1119 45F
1120 460
1121 461
1122 462
1123 463
1124 464
1125 465
1126 466
1127 467
1128 468
1129 469
1131 46B
1188 4A4
1189 4A5

Table 69. (Page 2 of 3) User Type Abend Codes

Syncsort MFX Programmer’s Guide 18–3

 User-Type Abend Codes Issued by MFX
_______________________________________
Decimal Hexadecimal
1190 4A6
1191 4A7
1192 4A8
1193 4A9
1194 4AA
1195 4AB
1197 4AD
1198 4AE
1812 714
1813 715
1814 716
1815 717
1816 718
1817 719
1818 71A
1819 71B
1820 71C
1837 72D
1838 72E
1981 7BD
1985 7C1
1986 7C2
1987 7C3
1990 7C6
1991 7C7
1992 7C8
1993 7C9
1994 7CA
1997 7CD
1998 7CE
2048 800
2049 801
2050 802
2081 821
2285 8ED
2320 910
2388 954
3585 E01

* The RC16=ABE option is specified and

there has been a critical error.
** The IOERR=ABE option is specified
and there has been an I/O error.
*** This is most commonly caused by the
release of MFX’s SVC not matching the
release of the MFX module.

Table 69. (Page 3 of 3) User Type Abend Codes

18–4 Syncsort MFX Programmer’s Guide

Before Contacting Precisely Support
All pertinent information (listings, dumps, maintenance level, SVC number, etc.)
should be available for easy reference when contacting Precisely Support. For error
conditions producing the WER999A message, the system dump and/or MFX SNAP
dump will prove helpful to an MFX analyst.

For other conditions cited with an “A” class message (e.g., WER039A
INSUFFICIENT VIRTUAL STORAGE), additional diagnostic information may be
required – a diagnostic SNAP dump can be produced by passing the DEBUG PARM
in the $ORTPARM DD statement or (for a JCL sort) in the // EXEC statement.
When using DEBUG, supply a SPYSET or SYSUDUMP DD statement to define an
appropriate SYSOUT data set for the dump. If the problem occurs in an application
using the OUTFIL OUTPUT feature, add the following DD statements in addition
to the DEBUG PARM:

//STDERR DD SYSOUT=*
//STDOUT DD SYSOUT=*
To get an accurate screen display of the maintenance level and SVC number, you
must determine whether the operating system found SYNCSORT in the link list
concatenation or found it in a STEPLIB concatenation. For link listed modules
issue the command TSO SYNCLEVL from an ISPF session active on the system
where the problem occurred. Otherwise, use the command TSO CALL
‘your.steplib(SYNCLEVL)’ substituting the library name that caused the problem.

Searching the Precisely Knowledge Base

The Precisely Knowledge Base contains numerous articles, including how-to’s and
resolutions to common problems. To help you research and resolve issues, the
Knowledge Base is continually updated to publicize issues that have already been
addressed by the Precisely support team. The Knowledge Base is available to
licensed users and is accessible through the Precisely Support website. Search on
the MFX product by release to find all relevant information about MFX.

Contacting Precisely Support

The Precisely Support portal is available to all registered users. To access technical
support for your location, visit https://www.precisely.com/support.

Syncsort MFX Programmer’s Guide 18–5

18–6 Syncsort MFX Programmer’s Guide
Index
Symbols

&DATE 2-168, 3-46 &TIME=(hp) 2-169

&DATENS=(xyz) 2-101, 2-108, 2-168 &TIMENS=(tt) 2-102, 2-110
&DATEx 2-34, 2-43–2-44 &YDDD 2-101, 2-109, 2-167, 2-170
&DATEx(c) 2-34, 2-43–2-44 &YDDDNS 2-101, 2-109, 2-168, 2-170
&DATExP 2-34, 2-43–2-44 $ORTPARM Statement 4-1, 4-12–4-15, 5-1, 6-2
&MULTIINDD 2-48, 2-52, 2-54–2-56, 2-135 for Century Window 4-14
&PAGE 2-103, 2-110, 3-46 with CENTWIN 4-14

AC Format 2-32, 2-67, 2-236 ASL Format 2-32, 2-67, 2-236

ACCEPT Parameter 2-90 Assembler Programs 1-1
ACS 2-251, 5-12 Invoking SyncSort from 6-1, 6-18
ALIGN Operator 13-17 AST Format 2-32, 2-67, 2-236
ALLDUPS 2-21, 2-24–2-26 ATOE 2-146
ALTSEQ 2-19–2-20, 6-9–6-10, 7-58 ATTACH Macro 6-3, 7-9, 7-30
AMODE 6-12 AUTHOR 2-127
AND Operator 2-36, 3-3 Authorization Messages 17-73
ANSI Control Characters 2-120–2-122 Averages
APPLICATION 2-127 See AVG
AQ Format 2-19, 2-32, 2-62, 2-67, 2-236 AVG 2-21, 2-23, 2-25–2-26, 2-114, 3-57

B Messages BNO 2-36

See BMSG Option BNZ 2-36
BACKGROUND COLOR 2-128 BO 2-35
BALANCE Option 5-2, 5-5, 14-4 BZ 2-36
BALN Option 6-7, 6-10 BKPTDSN Option (MAXSORT) 5-4, 9-11
BatchPipes/MVS 2-92, 4-5, 4-8–4-9 BLKCCH1 2-6, 2-97
BI Format 2-32, 2-62, 2-67, 2-236 BLKCCH2 2-6, 2-97
Binary Zeros, Insertion of 2-137–2-166, BLKCCT1 2-6, 2-98
3-29–3-31 BLKSIZE Parameter 4-4, 5-24, 5-27–5-28
BIT 2-147 Block List Exit Parameters 6-21
Bit Level Comparison 2-35 Block Size 5-27–5-28
Bit Level Logic 2-35–2-36, 2-49–2-51 BMSG Option 5-2, 5-5
Bit Level Processing BSAM 4-5, 4-9
BM 2-35 BUFOFF Parameter 4-4
BNM 2-36 BUILD 2-93–2-139, 2-203

C E15 7-19–7-28 C Programs 1-1

C E35 7-43–7-51 Century Window Processing 2-71, 2-162, 2-223,
C Exits 7-19–7-28, 7-43–7-51 2-226, 2-239–2-249

Syncsort MFX Programmer’s Guide Index–1

Century Window, with $ORTPARM 4-14 Concatenating Input Data Sets 4-6
CENTWIN Option 2-71, 2-239–2-249, 5-2, 5-6 VSAM files 4-6, 12-1
CENTWIN Processing, with OUTREC 2-223, COND Parameter (INCLUDE/OMIT) 2-31–2-53
2-226 Constant_name Statement 13-5
CENTWIN, with $ORTPARM 4-14 Contact details 18-6
CH Format 2-32, 2-62, 2-67, 2-236 Control Statement Syntax 2-14–2-17
CHANGE 2-182 Control Statements 2-1–2-257
Checkpoint-Restart 4-15, 14-7–14-10 See also $ORTPARM Statement
Automatic 14-8–14-9 See also Job Control Language
Deferred 14-9–14-10 See also SYSIN Statement
CKPT/CHKPT Parameter (MERGE) 2-78 ALTSEQ 2-19–2-20, 6-9–6-10, 7-58
CKPT/CHKPT Parameter (SORT) 2-250 Coding in Invoked Programs 6-2
CLC 9-15 Comments in 2-16
CLO Format 2-32, 2-67, 2-236 Continuation of 2-16–2-17
CMP 2-70, 2-238, 5-2, 5-8, 5-31, 9-15, 14-3 DEBUG 6-9
CMP Option 2-70, 5-2, 5-8, 14-3 Defaults 2-3–2-10
CMP=CLC 2-70, 2-239, 5-2, 5-8, 9-15, 14-3 DUPKEYS 1-5, 2-1, 2-3, 2-11, 2-21–2-27,
CMP=CPD 2-70, 2-238, 5-2, 5-8, 9-15, 14-3 4-2, 4-9, 5-13–5-14, 5-23, 5-33, 6-2,
COBOL E15 5-12, 7-11–7-19 6-7, 6-10, 7-28, 9-5, 10-3, 11-3, 14-2,
DATA DIVISION 7-14 14-4
ENVIRONMENT DIVISION 7-14 END 2-21, 2-28, 6-3
EXIT-STATUS Codes 7-15 for MAXSORT 9-10
Fixed-Length Records 7-12, 7-16–7-17 INCLUDE/OMIT 2-29–2-53, 3-3–3-6, 5-14,
IDENTIFICATION DIVISION 7-14 5-29, 6-9, 14-2
LINKAGE SECTION 7-11–7-14 INREC 2-54–2-56, 3-6–3-11, 6-9, 14-2
PROCEDURE DIVISION 7-14 JOIN 2-57
RETURN-CODE Codes 7-15 JOINKEYS 2-59
Variable-Length Records 7-13–7-14, Labels in 2-17
7-18–7-19 MERGE 2-54, 2-66–2-79, 5-12, 5-15, 5-28,
WORKING-STORAGE SECTION 7-14 6-3, 6-8, 7-57–7-58
COBOL E35 5-13, 7-32–7-42 MODS 2-80–2-83, 5-12, 6-3, 6-5, 6-9
DATA DIVISION 7-35 Notational Conventions 2-17
ENVIRONMENT DIVISION 7-35 OMIT 2-84
EXIT-STATUS Codes 7-36 OUTFIL 2-54, 2-85–2-133, 3-38–3-65, 6-2,
Fixed-Length Records 7-33, 7-38–7-39 6-9, 14-2
IDENTIFICATION DIVISION 7-35 OUTREC 2-54, 2-134–2-226, 3-25, 6-9,
LINKAGE SECTION 7-32–7-35
14-2
PROCEDURE DIVISION 7-35 Performance Considerations 14-2
Processing Sequence 2-12–2-14
RETURN CODE Codes 7-36–7-37
RECORD 2-227–2-230, 6-3
Variable-Length Records 7-34–7-35,
7-40–7-42 REFORMAT 2-231
WORKING STORAGE SECTION 7-35 Requirements for Disk Sort 2-11
COBOL Exits 7-5, 7-11–7-19, 7-32–7-42 Requirements for MAXSORT 2-11
COBOL Programs 1-1, 6-1, 14-2 Rules for Specifying 2-14–2-17
COBOL, and Century Window 4-14 SORT 2-54, 2-234, 5-12, 5-15, 5-28, 6-3,
CODE Parameter (ALTSEQ) 2-19
7-52, 7-57–7-58
Specifying Field Formats in 2-15
Coding Conventions 4-3–4-4
Specifying Field Lengths in 2-15
Collating Sequence 2-19–2-20, 2-67, 7-58
Specifying Field Positions in 2-15
Combining Records in a File 3-11
Specifying Parameters in 2-14–2-15
COMMAREA Option 5-2, 5-9, 7-5
SUM 2-54, 2-254–2-257, 3-11, 5-13–5-14,
Communication Area for Exits 5-2, 5-9, 7-5 6-9, 7-28, 14-2
Comparing Fields 2-31–2-53, 2-238, 3-3–3-11
Summary of Functions 2-1–2-2
Bit Level Criteria 2-35–2-36
Summary of Parameters and Defaults
Field to Constant Comparison 2-41 2-3–2-10
PD and ZD Field Comparison 2-70, 2-238 Use in Invoked Applications 6-3–6-4
Rules for Specifying Fields 2-70 CONVERT Parameter (OUTFIL) 2-95

Index–2 Syncsort MFX Programmer’s Guide

CONVERT Parameter (OUTREC) 2-192 Creating Input Data Sets for 4-5
Converting 2-162 Flow of 8-1–8-7
Converting Data 2-103, 2-111–2-112, 2-141, COPYALLOWED 2-128
2-147, 2-149, 2-163–2-177, 3-31–3-38, CORE Option 5-2, 5-10
5-6 COUNT 2-115–2-116
Format 2-103, 2-111, 2-115, 2-141–2-144 COUNT15 2-116
Converting Fixed-Length Records to Variable- CPU Option 5-2, 5-10, 14-4
Length Records 2-96 CRCX Option 6-7, 6-10
Converting SMF Formats 2-164 CSF Format 2-32, 2-68, 2-236
Converting Variable-Length Records 2-95, CSL Format 2-32, 2-67, 2-236
2-156, 2-192–3-38 CST Format 2-32, 2-67, 2-236
Converting Year Data 2-162 CTO Format 2-34, 2-69, 2-238
Copy 1-2 Cultural Environment 2-66, 2-234

DASD Data Set 4-4 See also Exit Programs, Link-editing

Data Set Placement 14-10 Coding Conventions 4-3–4-4
Data Utility Features 1-3, 3-1–3-65 JOBLIB Statement 4-4
Duplicate Records 3-11–3-13 Parameters
Examples, Index to 3-2 AMP 4-4
Input Record Selection 3-3–3-11 BLKSIZE 4-4
Input Records, Selection of Relevant Fields BUFND 4-4
3-6–3-11 BUFNI 4-4
Output Files, Multiple 3-63–3-65 BUFOFF 4-4
Output Records
Converting Data 3-34–3-36 BUFSP 4-4
DCB 4-4
Converting Data to Hexadecimal Format
3-33–3-34 DISP 4-4
Converting Data to Readable Form DSNAME/DSN 4-4
3-31–3-33 LABEL 4-4
Editing Data 3-34–3-36 LRECL 4-4
Editing of 3-25–3-38 OPTCD 4-4
Formatting Data Fields 3-36–3-38 RECFM 4-4
Inserting Binary Zeros 2-137–2-166, SPACE 4-4
3-29–3-31 UNIT 4-4
Inserting Blanks 3-27–3-29 VOLUME/VOL 4-4
Reordering Field Positions 3-26–3-29 SORTBKPT Statement 9-6–9-7
Output Reports SORTCKPT Statement 4-15
Counting Data Records 3-59–3-63 SORTIN Statement 4-5–4-7, 6-2
Headers and Trailers for 3-43–3-51 SORTINn Statement 4-7
SORTINnn Statement 6-2
Sectioning of 3-41–3-43
SORTJNF1 Statement 4-9
Totaling and Subtotaling Data
SORTJNF2 Statement 4-9
3-51–3-56 SORTMIn Statement 12-3
Sample Applications, Summarized 3-2 SORTMInn Statement 4-17, 12-2–12-4
DATE (&DATE) 2-100, 2-108–2-109, 2-168, SORTMODS Statement 4-15
3-46 SORTOFx Statement 4-9, 6-2
DATEADD 2-172, 2-174
SORTOFxx Statement 4-9, 6-2
DATEDIFF 2-174–2-175
SORTOU00 Statement 9-7–9-10
DB2 PARM 4-3, 5-5, 11-2, 11-4–11-8
SORTOUT Statement 4-9, 6-2
DC2 2-142
SORTPARn Statement 10-6–10-7
DCB Parameter 4-4, 5-3, 5-24, 5-27
SORTWKnn Statement 4-11, 6-2
DD Statements 4-4–4-17, 9-5–9-10, 10-3
SORTXDUP 4-9
$ORTPARM Statement 4-12–4-15, 6-2

Syncsort MFX Programmer’s Guide Index–3

 SORTXSUM 4-9 Records 2-97
STDERR Statement 4-16 Device Types 4-11
STDOUT Statement 4-16 DIAG Option 5-2, 5-11, 6-7, 6-10
STEPLIB Statement 4-4 Dictionary Feature 13-1
Summarized for Invoked Sort/Merge 6-2 Dictionary Statements 13-1–13-28
SYMNAMES Statement 13-1, 13-19, 13-26 Dictionary_names Statement 13-1, 13-4, 13-9,
SYMNOUT Statement 13-24 13-20, 13-23–13-25
SYSIN Statement 4-12 DISP Parameter 4-4, 5-27
SYSLIN Statement 4-16 DSN Parameter 4-4
SYSLMOD Statement 4-16 DSNAME Parameter 4-4
SYSOUT Statement 4-5, 6-2 DT 2-150–2-151
SYSPRINT Statement 4-16 DT1 2-164
ddname 4-1, 9-5, 10-3, 11-3 DT2 2-164
DEBUG Control Statement 6-9 DT3 2-164
DEBUG Option 5-2, 5-11, 18-1 DTNS Subparameter 2-150–2-152
Decimal Floating Point DUPKEYS 1-5, 2-1, 2-3, 2-11, 2-21–2-27, 4-2,
See FD Format 4-9, 5-13–5-14, 5-23, 5-33, 6-2, 6-7, 6-10,
Defining Output Data Sets 7-28, 9-5, 10-3, 11-3, 14-2, 14-4
See SORTOUT Files Duplicate Records 3-11–3-13
See SORTOUT Statement DYNALLOC Option 4-1, 5-2, 5-11–5-12
Defining Work Areas DYNALLOC Parameter (SORT) 2-250–2-251
See SORTWKnn Statement DYNATAPE Option (MAXSORT) 4-1, 5-4, 9-11,
Deleting Trailing Bytes From Fixed-Length 9-16–9-17

E10 2-82 Editing Masks 2-115, 2-179–2-182

E11 2-82, 7-6 ELAP Option 5-2, 5-13, 14-4
E14 4-11, 5-14, 7-7–7-8, 9-14 EMAIL 2-86, 2-125, 2-130–2-131, 3-66
E15 2-82, 5-14, 7-7–7-10, 7-59, 9-14, 15-5 E-mail Output File 2-86, 2-125, 2-130–2-131,
See also COBOL E15 3-66
E15 Option 5-2, 5-12, 7-11 END Control Statement 2-21, 2-28
E15=COB 5-2, 5-12 ENDREC Parameter (OUTFIL) 2-90–2-91
E16 4-11, 7-51, 9-14 EQ/OR Condition 2-36–2-37
E17 7-52 Equal-keyed Records 2-78, 2-251, 2-254, 5-13
E18 7-52–7-56 EQUALS Option 5-2, 14-3
E20 2-82 EQUALS/NOEQUALS Parameter (MERGE)
E21 2-82, 7-6 2-78
E25 7-7, 7-28–7-29, 9-14 EQUALS/NOEQUALS Parameter (SORT) 2-251
E27 7-52 Esoteric Names 10-7
E30 2-82 ETOA 2-146
E31 2-82, 7-6 ETOD 2-142, 2-164
E32 2-78, 6-2, 7-6 EXEC Statement 4-3, 9-4, 10-2, 11-2
E35 4-2, 7-7, 7-29–7-32, 7-59, 9-14 for DB2 Query 4-3
See also COBOL E35 for Disk Sort 4-3
E35 Option 5-2, 5-13 for MAXSORT 4-3, 9-4
E35=COB 5-2, 5-13 for MULTIIN 4-3
E37 7-52 for PARASORT 10-2, 11-2
E38 7-52–7-56 Exit Conventions 7-3
E39 7-52, 7-56–7-57 Exit Programs 2-80–2-83, 7-1–7-61
E61 2-82, 7-57–7-59, 9-14 Acting on Insufficient Storage 7-51
EDIT 2-115 Adding Records 7-29–7-32
Edit Patterns 2-177–2-182, 3-36 Analyzing Sort Input File 7-8–7-10
EDIT Subparameter 2-177 Changing Records 7-7–7-8, 7-28–7-32
Editing Data 2-157–2-186, 3-25–3-38 Checking Labels 7-52–7-56

Index–4 Syncsort MFX Programmer’s Guide

 Closing Exit Data Sets 7-52 Phases 7-1–7-2
Communication Area 7-5 Preparing for Other Exit Programs 7-6
Creating Input Records 7-6–7-10 Processing Read Errors 7-52–7-56
Creating Sort Input File 7-8–7-10 Processing Write Errors 7-52–7-56
Definition of 7-1 Program Labels 7-1
Deleting Records 7-7–7-8, 7-28–7-32 Register Conventions 7-4–7-5
End-of-File Routines 7-52–7-56 Revising Sort Input File 7-8–7-10
for Invoked Merge 7-6 Summary of Tasks Performed 7-2
Identified in MODS Control Statement 7-3 Summing Records 7-7–7-8, 7-28–7-29
Link-editing 7-3 VSAM Processing 7-52, 7-56
Link-editing at Execution Time with Disk Sort 2-82
DD Statements Required for 4-15–4-17 with MAXSORT 2-82, 9-14
Loading into Main Storage 7-3 with PARASORT 2-82
MAXSORT 9-15 Exit Programs, Link-editing 4-17
Modifying Collating Sequence 7-57–7-59 Exit-Name Parameter (MODS) 2-80–2-82
EXTCOUNT Option 5-2, 5-14

F1 Parameter 2-61 2-159–2-161, 2-237, 13-13

F1 Parameter (JOIN) 2-57 TS 2-32, 2-67, 2-236
F2 Parameter 2-61 UFF 2-33, 2-68, 2-114, 2-142, 2-157,
F2 Parameter (JOIN) 2-57 2-159–2-161, 2-237, 13-13
FD Format 2-23, 2-29, 2-47–2-48, 2-68, 2-103, Y2B 2-33, 2-68, 2-71, 2-162, 2-164, 2-237,
2-111–2-113, 2-115, 2-143, 2-159, 2-161, 2-239
2-179, 2-236, 2-255 Y2C 2-33, 2-68, 2-71, 2-162, 2-164, 2-237,
FI Format 2-32, 2-62, 2-67, 2-236 2-240
Field Format Codes 2-32, 2-67–2-69, Y2D 2-33, 2-68, 2-72, 2-163–2-164, 2-237,
2-235–2-238 2-240
AC 2-32, 2-67, 2-236 Y2P 2-33, 2-69, 2-72, 2-163–2-164, 2-237,
AQ 2-19, 2-32, 2-62, 2-67, 2-236 2-240
ASL 2-32, 2-67, 2-236 Y2S 2-33, 2-44, 2-69, 2-73, 2-162, 2-164,
AST 2-32, 2-67, 2-236 2-237, 2-241
BI 2-32, 2-62, 2-67, 2-236 Y2T 2-33, 2-45, 2-69, 2-163, 2-237, 2-239
CH 2-32, 2-62, 2-67, 2-236 Y2U 2-33, 2-45, 2-69, 2-163, 2-237, 2-239
CLO 2-32, 2-67, 2-236 Y2V 2-33, 2-45, 2-69, 2-163, 2-237, 2-239
CSF 2-32, 2-68, 2-236 Y2W 2-33, 2-45, 2-69, 2-163, 2-237, 2-239
CSL 2-32, 2-67, 2-236 Y2X 2-33, 2-45, 2-69, 2-163, 2-237, 2-239
CST 2-32, 2-67, 2-236 Y2Y 2-33, 2-45, 2-69, 2-163, 2-237, 2-239
CTO 2-34, 2-69, 2-238 Y2Z 2-34, 2-44, 2-69, 2-71, 2-162, 2-164,
FD 2-23, 2-29, 2-47–2-48, 2-68, 2-103, 2-237, 2-240
2-111–2-113, 2-115, 2-143, 2-159, ZD 2-34, 2-62, 2-69, 2-164, 2-238
2-161, 2-179, 2-236, 2-255 Field_name Statement 13-9–13-10
FI 2-32, 2-62, 2-67, 2-236 Fields 2-134–2-226
FL 2-32, 2-67, 2-236 Binary 2-29, 2-35–2-36, 2-235
FS 2-32, 2-68, 2-236 Bit Level Comparison 2-35
List of Valid Formats 2-32, 2-67–2-69, Comparison of 2-31–2-53, 2-238, 3-3–3-11
2-236–2-238 Format 2-32, 2-64, 2-67–2-160,
LS 2-32, 2-68, 2-236 2-235–2-238
OL 2-32, 2-67, 2-236 Insertion of Binary Zeros 3-29–3-31
OT 2-34, 2-69, 2-238 Insertion of Blanks 3-27–3-29
PD 2-32, 2-62, 2-67, 2-236, 2-238 Length 2-235
PD0 2-32, 2-68, 2-73, 2-164, 2-236, 2-238, Position 2-147, 2-149, 2-163, 2-235, 5-6
2-243 Reordering 3-26–3-29
SFF 2-33, 2-68, 2-114, 2-142, 2-157, Rules for Specifying 2-70, 2-238

Syncsort MFX Programmer’s Guide Index–5

 Selection of 3-6–3-11 FILSZ Option 5-2, 5-14–5-15
Specifying Format 2-15 FILSZ Parameter (SORT) 2-252
Specifying Length 2-15, 3-6 FINDREP Parameter 2-94, 2-137,
Specifying Position 2-15, 3-6 2-193–2-194, 2-196–2-199, 2-204, 17-51
Substring Comparison 2-46 FIRSTDUP 2-21, 2-24–2-26
Pattern Constant (Wildcard) 2-31, 2-46, Fixed-Length Records 2-15, 2-70, 2-156, 2-192,
2-51–2-52 2-227, 7-12, 7-16–7-17
FIELDS Parameter (INREC) 2-55 Maximum Length 4-5, 12-3
FIELDS Parameter (JOINKEYS) 2-61 FL Format 2-32, 2-67, 2-236
FIELDS Parameter (MERGE) 2-66–2-70 FLAG Option 5-2, 5-15, 6-10
FIELDS Parameter (OUTREC) 2-137–2-186 Flow of the Sort 8-1–8-7
FIELDS Parameter (SORT) 2-234–2-249 FNAMES Parameter (OUTFIL) 2-89
FIELDS Parameter (SUM) 2-254–2-255 fo 2-103, 2-111, 2-115, 2-143, 13-23
FIELDS Subparameters 2-137 FONT 2-128–2-129, 3-66
FIELDS=COPY 2-70, 2-79, 2-249 FONTHn 2-128, 3-66
FIELDS=NONE 2-21, 2-24, 2-254 FONTTn 2-128, 3-66
FILE Parameter (JOINKEYS) 2-60 FORMAT 2-31, 2-64, 2-66–2-67, 2-235, 2-253,
File Size 2-255, 2-257
See FILSZ Option FORTRAN Programs 1-1, 6-1
FILES Parameter (MERGE) 2-78 FS Format 2-32, 2-68, 2-236
FILES Parameter (OUTFIL) 2-88, 3-63–3-65 FTOV Parameter (OUTFIL) 2-96
FILL Parameter (REFORMAT) 2-232 Full-date formats 2-74, 2-244

Generating Run-Time Constants 2-166 2-174–2-175

GETMAIN 2-81, 6-3 Gregorian Date Output Format 2-155
Gregorian 2-134, 2-138, 2-150–2-153, Group, OUTFIL 2-88–2-89

HBSI Option 4-6, 5-2, 5-15 HISTE15 7-8, 15-5–15-7, 15-10, 15-13
HBSO Option 4-10, 5-2, 5-16 HISTOGRM 2-229–2-230, 5-17, 7-8,
HEADER1 / HEADER2 Parameter (OUTFIL) 15-1–15-14
2-98, 3-43 Control Parameters 15-2–15-4
HEADER3 2-118, 3-46–3-49 Executing through E15 15-5–15-7
HEX 2-139, 2-145–2-146, 2-156, 3-33 Job Control Language 15-5
HFS 4-5, 4-8–4-9 Messages 15-10–15-14
Hiperbatch 5-16 Output Samples 15-8
Hiperbatch Processing 5-15–5-16 HTML Output File 2-86, 2-125–2-126, 2-129,
3-66

IFOUTLEN 2-94, 2-208 Incore Sort 2-78, 2-250, 2-256, 8-1, 14-4
IFTHEN 2-94, 2-198–2-208, 2-217 Input Data Sets
IFTRAIL Parameter 2-122 Concatenation of 4-6
INCLUDE/OMIT Control Statement 2-29–2-53, Input Records
3-3–3-6, 5-14, 5-29, 6-9, 14-2 Selection of 3-3–3-11
INCLUDE/OMIT Parameter (JOINKEYS) 2-64 INREC Control Statement 2-54–2-56, 3-6–3-11,
INCLUDE/OMIT Parameter (OUTFIL) 6-9, 14-2
2-89–2-90 Installation Options

Index–6 Syncsort MFX Programmer’s Guide

 Precedence Rules 5-1 LOAD 6-2
Intermediate Storage 4-11 XCTL 6-2
Invoking SyncSort from a Program 6-1–6-18 MAXSORT 9-15
Assembler Programs 6-2 Performance Considerations 6-1
DD Statements 6-2 Restrictions on Control Statements 6-3–6-4
Macro Instructions for Invoking SyncSort from IO Option 5-2–5-3, 5-16, 14-4
an Assembler Program IOERR Option 5-2, 5-16, 6-10
ATTACH 6-2 IOERR=ABE 5-3
LINK 6-2

JCL for DB2 Query 11-2

See Job Control Language for HISTOGRM 15-5
JCL and Required Control Statements 3-67 for MAXSORT 9-4, 9-15
JFY 2-145, 2-186–2-189 for MULTIIN 12-2
Job Control Language 4-1, 4-17 for PARASORT 10-2
Control Statement Examples Initiating SyncSort 4-12, 6-1
Copy without Exit Routines 4-24–4-25 JOB Statement 4-1
Merge without Exit Routines 4-22–4-23 Sorts without Exit Routines 4-18
Multiple Output Files 4-29, 4-31 JOBLIB Statement 4-4
Sort with Exit Routine to be Link-edited JOIN
4-27–4-29 Control Statement 2-57
Sort with Link-edited Exit Routine User Guide 1-10
4-25–4-26 JOINKEYS 2-59
Sorts without Exit Routines 4-22 JPn 5-3, 5-16
Julian 2-134, 2-151–2-153, 2-172–2-173,
DD Statement 4-1 2-175
Elements of 4-1 Julian Date Output Formats 2-154
EXEC Statement 4-1, 5-1, 6-1

KEY Messages 17-73 KEYWORDS 2-127

L6 Option 5-3, 5-17, 15-1 lowercase to uppercase 2-145, 2-223

L7 Option 5-3, 5-17, 15-1 LRECL 2-104, 2-117, 2-120, 2-228–2-229, 4-4,
LABEL Parameter 4-4, 5-27 5-24, 5-27–5-28
LANDSCAPE 2-127 LS Format 2-32, 2-68, 2-236
Language Environment for z/OS 5-18 LTOU 2-146, 2-223
LASTDUP 2-21, 2-24–2-26
LENGTH 2-115
LENGTH Parameter (RECORD) 2-227–2-229
LENGTH Subparameter 2-178
LINES Parameter (OUTFIL) 2-119–2-122
LINK Macro 6-3, 7-9, 7-30
LIST Option 5-2, 5-17, 6-7, 6-10
LOAD Macro 6-3
LOCALE Option 5-3, 5-18–5-19, 7-58
Processing with INCLUDE/OMIT 2-29

Syncsort MFX Programmer’s Guide Index–7

M

Macro Instructions (Assembler) 6-2–6-3 Tuning 9-23

ATTACH Macro 6-2 User Guide 1-10
LINK Macro 6-2 MAXSORT Option (MAXSORT) 9-11
LOAD Macro 6-2 MAXWKSP Option (MAXSORT) 5-4, 9-12
XCTL Macro 6-2 Merge
MARGINS 2-127 Creating Input Data Sets 4-7
Masks 2-115 Creating Input Records for Invoked Merge
MAX 2-114 7-6
Maximums, Example 3-57 Flow of 8-1–8-7
MAXSORT 4-1, 4-7, 9-1–9-23 MERGE Control Statement 2-54, 2-66–2-79,
DD Statements for 4-17, 9-5–9-10 5-12, 5-15, 5-28, 6-3, 6-8, 7-6, 7-57–7-58
EXEC Statement for 4-3, 9-4 Merging 1-2
Exit Programs 9-14–9-15 Phases of 1-2
File Size 4-7 Message Data Set 5-21–5-24
Invoking from a Program 9-15 Messages 4-5, 5-5, 5-15, 5-20–5-21, 17-1
JCL/Control Stream Examples 9-18–9-23 See also FLAG Option
Operator Interface 9-16–9-17 Authorization Messages 17-73
PARM Options 4-3, 5-4, 9-11–9-14, 11-4 HISTOGRM 15-10–15-14
BKPTDSN 5-4, 9-11 KEY Messages 17-73
DYNATAPE 5-4, 9-11, 9-16–9-17 PROC SYNCSORT Messages 17-72–17-73
MAXSORT 9-11 MIN 2-113
Minimums, Example 3-57
MAXWKSP 5-4, 9-12
MINWKSP Option (MAXSORT) 5-4, 9-12
MINWKSP 5-4, 9-12
Missing Field Bytes, Record 2-95
NODYNATAPE 5-4, 9-11 Mm Subparameter 2-179
RESTART 5-4, 9-13 MODS Control Statement 2-80–2-83, 5-12, 6-3,
SORTSIZE 5-4, 9-13 6-5, 6-9, 7-3, 7-6, 7-9, 7-30
SORTTIME 5-4, 9-13 MSG Option 5-3, 5-20–5-21, 6-10
TAPENAME 5-4, 9-14 MSGDD Option 5-3, 5-21, 7-11, 7-32
Performance Considerations 9-23, 14-1 MULTIFETCH PARM 11-4
PGM Names 4-3 MULTIIN 2-48, 2-52, 2-55, 2-135, 4-3, 4-5–4-6,
Starting 9-15 4-17, 12-1–12-2
Starting through JCL 9-4, 9-15 Multiple Lines, Record 2-94
Summary of Control Statements for 2-11 Multiple Output 2-88–2-89, 3-63–3-65
Multiple Output Files 2-85

National Language 2-29, 2-66, 2-234, 5-18 NOTMTOFL 2-124

with INCLUDE/OMIT 2-29 NOTMTOUT 5-3, 5-21
NE/AND Condition 2-36–2-37 NULLOFL Parameter (OUTFIL) 2-7, 2-85,
NOCOMMAREA Option 5-2, 5-9 2-87, 2-124
NODETAIL Parameter (OUTFIL) 2-123 NULLOUT Option 5-3, 5-22
NODUPS 2-21, 2-24–2-26 NUM 2-47–2-48, 2-52
NODYNATAPE Option (MAXSORT) 5-4, 9-11 NZDPRINT 5-4
NOEQUALS Option 5-2 NZDPRINT Option 5-33
NOIOERR Option 5-2, 5-16
NOLIST Option 5-2, 5-17
NORC16 Option 5-3
NORESET Option 5-3, 5-26
NORLSOUT Option 5-3, 5-26
NOSEQCK Parameter 2-63
Notational Conventions 2-17

Index–8 Syncsort MFX Programmer’s Guide

O

OL Format 2-32, 2-67, 2-236 Editing Data 3-34–3-38

OMIT Control Statement Formatting 2-93–2-94, 2-134–2-226,
See INCLUDE/OMIT 3-25–3-38
ONLY Parameter (JOIN) 2-58 Output Reports 2-85–2-133, 3-41–3-65
Operator 2-141 Averaging Data 3-57
Operator Interface 9-16–9-17 Ending Record Number 2-90–2-91
Operator Statement 13-15 Headers 2-98, 2-119–2-122, 3-43–3-51
OPTCD Parameter 4-4 Obtaining maximums 3-57
OPTELAP Option 5-3, 5-22 Obtaining minimums 3-57
OR Operator 2-36, 3-3 Pages, Logical 2-119–2-122
OSCL Option 6-7, 6-10 Records Included 2-90–2-91
OT Format 2-34, 2-69, 2-238 Saving Records 2-91
OUTFIL Control Statement 2-54, 2-85–2-133, Sections 2-118–2-119, 3-41–3-43
3-38–3-65, 6-2, 6-9, 14-2 Starting Record Number 2-90
OUTFIL Group 2-88–2-89 Subtotaling Data 3-51–3-56
Output Data Sets 4-9–4-10 Totaling Data 3-51–3-56
See SORTOUT Files Trailers 2-105, 2-119–2-122, 3-43–3-63
See SORTOUT Statement Output Space
Output Files, Multiple 2-85, 3-63–3-65, See Secondary Allocation
4-29–4-31 OUTREC Control Statement 2-54, 2-134–2-226,
Output Lines, Multiple 2-94, 2-99, 3-39 3-25, 6-9, 14-2
OUTPUT Parameter (OUTFIL) 2-125 OUTREC Parameter (OUTFIL) 2-93–2-94
Output Records OVERLAY 2-208–2-209, 2-211, 2-217
Converting Data 3-31–3-38 OVFLO Option 5-3, 5-23
Distributing 2-91 OWNERPASSWORD 2-128

PAD Option 5-3, 5-23 EQUALS/NOEQUALS 2-78, 2-251

PAGE (&PAGE) 2-103, 2-110, 3-46 Exit-Name 2-80–2-82
PAGEHEAD 2-119–2-120 F1 2-57
Pages, Defining Logical 2-119–2-122 F2 2-57
PAGESIZE 2-127 FIELDS 2-61, 2-66–2-70, 2-137–2-186,
Parameter List 2-234–2-249, 2-254–2-255
24-bit 6-4–6-12, 6-18, 7-3–7-4, 7-30 FIELDS=COPY 2-21, 2-70, 2-249, 2-254
Optional Parameters 6-8 FIELDS=NONE 2-24–2-25, 2-254, 3-13
31-bit 6-12–6-18, 7-4, 7-10, 7-31 FILE 2-60
32-bit 7-4–7-10, 7-30–7-31 FILES 2-78, 2-88, 3-63–3-65
64-bit 6-18–6-21, 7-4–7-10, 7-30–7-31 FILL 2-232
Parameters (Control Statements) FILSZ 2-252
ALLDUPS 2-24–2-26 FIRSTDUP 2-24–2-26
BLKCCH1 2-97 FNAMES 2-89
BLKCCH2 2-97 FORMAT 2-31, 2-64, 2-67, 2-234–2-235,
BLKCCT1 2-98 2-253, 2-255, 2-257
BUILD 2-93, 2-137, 2-196, 2-199, 2-203, FTOV 2-96
2-209, 2-211 HEADER 14-2
CENTWIN 2-71, 2-239 HEADER1/HEADER2 2-98, 3-43
CKPT/CHKPT 2-78, 2-250 IFTHEN 2-94, 2-137, 2-196, 2-198–2-200,
CODE 2-19 2-203–2-204, 2-207–2-208
COND 2-31–2-53 INCLUDE/OMIT 2-64, 2-89–2-90
CONVERT 2-95, 2-192–3-38 LASTDUP 2-24–2-26
DYNALLOC 2-250–2-251 LENGTH 2-227–2-229
ENDREC 2-90–2-91 LINES 2-119–2-122

Syncsort MFX Programmer’s Guide Index–9

 NODETAIL 2-123 PARM Options 5-1
NODUPS 2-25–2-26 &MULTIINDD 2-48, 2-52, 2-54–2-56,
NOTMTOFL 2-124 2-135
NULLOFL 2-7, 2-85, 2-87, 2-124 See also $ORTPARM Statement
ONLY 2-58 BALANCE 5-2, 14-4
OUTPUT 2-125–2-126, 3-66 BALN 6-7, 6-10
OUTREC 2-93–2-94, 14-2 BKPTDSN 5-4
OVERLAY 2-95, 2-137, 2-196, 2-199, BMSG 5-2, 5-5
2-203, 2-208–2-209, 2-211 CENTWIN 2-70–2-71, 2-239–2-249, 5-2,
PARSE 2-94, 2-137, 2-199, 2-204 5-6
REMOVECC 2-124 CMP 2-70, 2-238, 5-2, 5-8, 9-15, 14-3
REPEAT 2-90 CMP=CLC 2-70, 2-239, 5-2, 5-8, 9-15, 14-3
SAMPLE 2-91 CMP=CPD 2-70, 2-238, 5-2, 5-8, 9-15, 14-3
SAVE 2-91 COMMAREA 5-9, 7-5
SECTIONS 2-118–2-119, 3-41, 14-2 CORE 5-10
SIZE 2-229, 2-252 CPU 5-2, 14-4
SKIPREC 2-79, 2-252, 14-2 CRCX 6-7, 6-10
SORTED 2-63 DB2 5-5, 11-4
SPLIT 2-91–2-92 DEBUG 5-11
SPLIT1R 2-7, 2-92–2-93, 8-5–8-7, 17-51 DIAG 5-11, 6-7, 6-10
SPLITBY 2-92 DYNALLOC 4-1, 5-11–5-12
STARTREC 2-90 DYNATAPE 4-1, 5-4
STOPAFT 2-79, 2-252, 14-2 E15 5-12, 7-11
TRAILER 14-2 E35 5-13
TRAILER1/TRAILER2 2-105, 3-43, 3-59 ELAP 5-2, 5-13, 14-4
TYPE 2-64, 2-227 EQUALS 5-2, 14-3
UNPAIRED 2-57 EXTCOUNT 5-2, 5-14
VLFILL 2-95 FILSZ 5-2, 5-14–5-15
VLTRIM 2-97 FLAG 5-2, 5-15, 6-10
XDUP 2-25 for Disk Sort 5-1
XSUM 2-255 for MAXSORT 5-4, 9-11–9-14, 11-4
Parameters (DD Statements) for PARASORT 5-5
AMP 4-4 Format of 5-1
BLKSIZE 4-4, 5-24, 5-27–5-28 HBSI 4-6, 5-2, 5-15
BUFND 4-4 HBSO 4-10, 5-2, 5-16
BUFNI 4-4 IO 5-2–5-3, 5-16, 14-4
BUFOFF 4-4 IOERR 5-2, 5-16, 6-10
BUFSP 4-4 L6 5-17, 15-1
DCB 4-4, 5-3, 5-24, 5-27 L7 5-2, 15-1
DISP 4-4, 5-27 LIST 5-2, 5-17, 6-7, 6-10
DSNAME/DSN 4-4 LOCALE 5-3, 7-58
LABEL 4-4, 5-27 MAXWKSP 5-4
LRECL 4-4, 5-24, 5-27–5-28 MINWKSP 5-4
OPTCD 4-4 MSG 5-2, 5-20–5-21, 6-10
RECFM 4-4, 5-24, 5-27–5-28 MSGDD 5-2, 5-21, 7-11, 7-32
SPACE 4-4, 5-3, 5-25 MULTIFETCH 11-4
UNIT 4-4 MULTIIN 2-52, 2-55, 2-135, 4-3, 4-17
VOLUME/VOL 4-4 NOCOMMAREA 5-9
PARASORT 4-1, 4-7, 10-1–10-9 NODYNATAPE 5-4
DD Statements for 4-17, 10-3 NOEQUALS 5-2
EXEC Statement for 10-2, 11-2 NOIOERR 5-2, 5-16
JCL for 10-2 NOLIST 5-2, 5-17
JCL/Control Stream Examples 10-9 NORC16 5-3
PARM Options 4-3, 5-5 NORESET 5-3, 5-26
PGM Names 4-3 NORLSOUT 5-3, 5-26
SORTIN Statement for 10-4 NOTMTOUT 5-3, 5-21

Index–10 Syncsort MFX Programmer’s Guide

 NULLOUT 5-22 ZIIPPCT 5-33
NZDPRINT 5-4, 5-33 PARSE 2-209–2-219
OPTELAP 5-22 Pattern Constant (Wildcard) 2-31, 2-46,
OSCL 6-7, 6-10 2-51–2-52
OVFLO 5-23 PD Format 2-32, 2-62, 2-67, 2-70, 2-236, 2-238
PAD 5-23 PD0 Format 2-32, 2-68, 2-73, 2-164, 2-236,
PEER 6-7, 6-10 2-238, 2-243
POLY 6-7, 6-10 PDC Output Field Format 2-143, 2-159, 13-22,
Precedence Rules 5-1 13-24
PRINT121 5-3 PDF Output Field Format 2-144, 2-159, 13-22,
RC16 6-10 13-24
RELEASE 5-3 PDF Output File 2-86, 2-125–2-126,
RESERVE 5-3, 5-25
2-128–2-129, 3-66, 3-68
PEER Option 6-7, 6-10
RESERVEX 5-3, 5-25
RESERVEZ 5-26 Performance Considerations 14-1
Control of System Resource Usage 14-2
RESET 5-3, 5-26
Control Statements 14-2
RESTART 5-4
JCL- vs. Program-Invoked Sort 14-2
RLSOUT 5-3, 5-26
Memory Management 14-2
SDB 5-3, 5-27–5-28
PARM Options 14-3
SIZE 5-10
Phases of Copying, Merging, Sorting 8-1–8-7
SKIPREC 5-3, 7-52, 14-3
SORTSIZE 5-4 PipeSort 1-7, 2-85, 3-63, 16-2
PL/1 Programs 1-1, 6-1
SORTTIME 5-4
POLY Option 6-7, 6-10
Specification in JCL-initiated Applications
5-1 PORTRAIT 2-127
Specification in Program-initiated POSITION Operator 13-15–13-16
Applications 5-1 Precedence Rules 5-1
STOPAFT 5-3, 5-29, 14-3 Precisely
Summarized 5-2 Contacting for support 18-6
SWCRYPT 5-29 PRINT121 Option 5-3, 5-24
TAPENAME 5-4 PRINTINGALLOWED 2-128
TRUNC 5-29 PROC SYNCSORT - An Accelerator for SAS™
UNINTDS 5-4, 5-30 Sorting 1-7
VLTEST 5-3, 5-30 PROC SYNCSORT Messages 17-72–17-73
VLTESTI 5-4, 5-32 PROCSort 16-1
VSAMEMT 5-4, 5-33 Program Exits
ZDPRINT 5-4, 5-33 See Exit Programs

RC16 Option 5-3, 5-23 REFORMAT Control Statement 2-231

RDW 2-15, 2-228 Reformatting Records 2-54–2-56, 2-93–2-94,
RECFM Parameter 4-4, 5-24, 5-27–5-28 2-134–2-226, 3-26–3-29
RECORD Control Statement 2-227–2-230, 6-3 CHANGE Subparameter 2-182
Record Counts 3-59–3-63 Column Alignment 2-99
Record Format 2-227–2-230 Data Conversion and Editing 2-157–2-223,
Record Length 2-227–2-230 3-31–3-38
Maximum for Fixed-Length Records 4-5, Inserting Binary Zeros 2-166, 3-29
12-3 Inserting Blanks or Spaces 2-166
Maximum for Variable-Length Records 4-5, Inserting Hexadecimal Characters 2-166
12-3 Inserting Literal Characters 2-166
Record Selection 2-29–2-53, 2-79, 2-89–2-90, Missing Field Bytes 2-95
2-252, 3-3–5-29 OUTREC Parameter 2-93–2-94
Using Bit-Level Logic 2-49–2-51 Record on Multiple Lines 2-94
Records, Distributing Output 2-91 Replace 2-182

Syncsort MFX Programmer’s Guide Index–11

 Search and Replace 2-182 RESERVEZ Option 5-3, 5-26
Variable-Length Records 2-156 RESET Option 5-3, 5-26
Register Conventions 7-4–7-5 RESTART Option (JCL) 14-9
Registers RESTART Option (MAXSORT) 5-4, 9-10, 9-13,
See Register Conventions 9-21
RELEASE Option 5-3, 5-24 RESTART Option (OUTREC) 2-171–2-172
REMOVECC Parameter (OUTFIL) 2-124 RETRY 2-251, 5-12
REPEAT Parameter (OUTFIL) 2-90 Return Codes 6-10, 6-15
Report Writing 2-85, 2-98–2-133, 3-41–3-65 REXX Exits 2-81, 7-59–7-61
Repositioning Record Fields 3-26–3-29 REXX Programs 1-1
RESERVE Option 5-3, 5-25 RLSOUT Option 5-3, 5-26
Reserved Words 13-3 RTF Output File 2-86, 2-125–2-126, 2-129,
RESERVEX Option 5-3, 5-25–5-26 3-66
Run-Time Constants, Generating 2-166

SAMPLE Parameter (OUTFIL) 2-91 SORTIN Statement 4-1, 4-5–4-7, 5-27, 6-2, 7-9,
SAVE Parameter (OUTFIL) 2-91 10-4
SC 2-251, 5-12 for PARASORT 10-4
SDB 4-19–4-21, 4-24, 4-28, 5-27 Sorting
SDB Option 5-3, 5-27–5-28 Phases of 1-2
Search and Replace 2-182 Types of 1-2
Secondary Allocation 4-10–4-11 Sorting Technique
SECTIONS Parameter (OUTFIL) 2-118–2-119, Disk Sort 4-1, 14-1
3-41 MAXSORT 4-1, 4-7, 4-17, 9-1–9-23, 14-1
SEQNUM 2-54, 2-171 PARASORT 4-1, 4-17, 10-1–10-9
SFF Format 2-33, 2-68, 2-114, 2-142, 2-157, Performance Considerations 14-1
2-159–2-161, 2-237, 13-13 SORTINn Statement 4-1, 4-7
SIGNS 2-115 SORTINnn File 5-13, 7-7
SIZE Option 5-2, 5-10 SORTINnn Statement 4-1, 4-7, 6-2, 7-6
SIZE Parameter (SORT) 2-229, 2-252 SORTJNF1 Statement 4-9
SKIP 2-118 SORTJNF2 Statement 4-9
SKIP Operator 13-17, 13-26 SORTMIn Statement 2-48, 2-55, 12-3
SKIP Parameter (OUTFIL) 2-118 SORTMInn Statement 2-48, 2-55, 4-5, 4-17,
SKIPREC Option 5-3, 5-28, 7-52, 14-3 12-3
SKIPREC Parameter (MERGE) 2-79 SORTMODS Library 7-3
SKIPREC Parameter (SORT) 2-252 SORTMODS Statement 4-1, 4-15
SMF Date and Time Formats 2-164 SORTOFx File 5-26–5-28
SNAP Dump 5-11 SORTOFx Statement 4-1, 4-9, 6-2
Sort SORTOFxx File 5-4, 5-26–5-27, 7-31, 7-36
Creating Input Data Sets for 4-5 SORTOFxx Statement 4-1, 4-9, 6-2
Flow of 8-1–8-7 SORTOU00 Statement 9-7–9-10
SORT Control Statement 2-54, 2-234–2-253, SORTOUT File 4-1, 4-9, 5-3–5-4, 5-16,
5-12, 5-15, 5-28, 6-3, 7-52, 7-57–7-58 5-26–5-28, 7-31, 7-36, 7-52
SORT/MERGE 6-2–6-4 SORTOUT Files 6-2
SORT/MERGE Options SORTOUT Space 5-26
Precedence Rules 5-1 SORTOUT Statement 4-1, 4-9, 6-2, 7-30
SORTBKPT Statement 9-6–9-7 SORTOUT VSAM File 5-26
SORTCKPT Statement 4-1, 4-15 SORTOUT Write Errors 7-56
SORTED Parameter (JOINKEYS) 2-63 SORTPARn Statement 10-6–10-7
SORTIN End-of-File 7-10, 7-15 SORTSIZE Option (MAXSORT) 5-4, 9-13
SORTIN File 4-1, 4-5, 5-3, 5-15, 6-8, 6-12, 6-17, SORTTIME Option (MAXSORT) 5-4, 9-13
7-8, 7-10, 7-12–7-15, 7-51 SORTWK
SORTIN Processing 7-52, 7-54 Dynamic Allocation of 5-11

Index–12 Syncsort MFX Programmer’s Guide

SORTWKnn File 7-7 SUBCOUNT=(...) 2-116
SORTWKnn Statement 4-1, 4-11, 6-2 SUBCOUNT15 2-117
See also DYNALLOC Option SUBJECT 2-127, 2-131, 3-66
Conditions of Use 4-11–4-12 SUBMAX 2-114
SortWriter 2-85, 2-98 SUBMIN 2-114
Sample Report 1-5 Substring Comparison 2-46
SORTXDUP 2-21, 2-25, 4-2, 4-9, 6-2, 9-5, 10-3, Pattern Constant (Wildcard) 2-31, 2-46,
11-3 2-51–2-52
SORTXSUM 4-9, 5-26, 6-2 SUBTOTAL 2-113
SPACE Parameter 4-4, 5-3, 5-25 SUM Control Statement 2-54, 2-254–2-257,
Special Esoteric Names 10-7 3-11, 5-13–5-14, 6-9, 7-28, 14-2
SPLIT Parameter (OUTFIL) 2-91–2-92 SWCRYPT Option 5-4, 5-29
SPLIT1R 2-7, 2-92–2-93, 8-5–8-7, 17-51 Symbols 13-1
SPLITBY Parameter (OUTFIL) 2-92 SYMNAMES Statement 13-1, 13-19, 13-26
SQZ 2-145–2-192 SYMNOUT Statement 13-24
STARTREC Parameter (OUTFIL) 2-90 SyncSort
STDERR Statement 2-125, 4-2, 4-16 Data Utility 1-3, 3-1–3-65
STDOUT Statement 2-125, 4-2, 4-16 Description of 1-1
STEPLIB Statement 4-4 Features of 1-1–1-6
STOPAFT Option 5-3, 5-29, 14-3 Initiation of 1-1
STOPAFT Parameter 2-65 SortWriter 1-5
STOPAFT Parameter (MERGE) 2-79 SYSIN Statement 4-1, 4-12
STOPAFT Parameter (SORT) 2-252 SYSLIN Statement 4-1, 4-16
Storage SYSLMOD Statement 4-1, 4-16
Disk 4-11 SYSOUT
Intermediate 4-11 See Message Data Set
SUB SYSOUT Statement 4-1, 4-5, 6-2
See SUBTOTAL SYSPRINT Statement 4-1, 4-16
SUBAVG 2-114 System Abend (0C7) 5-8, 5-31
SUBCOUNT 2-116 System Symbols 13-7

TAPENAME Option (MAXSORT) 5-4, 9-14 TM1 2-164

TASKID Parameter 2-64 TM2 2-164
Technical support TM3 2-164
Contacting Precisely 18-6 TM4 2-164
The 64-Bit Parameter List 6-18 TO=fo 2-103, 2-111, 2-115, 2-143, 13-23
TIME (&TIME) 2-102, 2-109, 2-167 TOD 2-142, 2-164
Time of Day Formats 2-164 TOGREG 2-153
DC1 2-142, 2-165 TOJUL 2-153–2-154
DC2 2-142, 2-165 TOT
DC3 2-142, 2-165 See TOTAL
DE1 2-142, 2-165 TOTAL 2-112
DE2 2-142, 2-165 TOTAL/TOT 2-112
DE3 2-142, 2-165 TRAILER1/TRAILER2 Parameter (OUTFIL)
TC1 2-142, 2-165 2-105, 3-43, 3-59
TC2 2-142, 2-165 TRAILER3 2-105, 2-116, 2-118–2-120, 3-56,
TC3 2-142, 2-165 3-59, 3-62–3-63
TC4 2-142, 2-165 TRAN subparameter 2-145–2-146
TE1 2-142, 2-165 TRUNC Option 5-4, 5-29
TE2 2-142, 2-165 TS Format 2-32, 2-67, 2-236
TE3 2-142, 2-165 TYPE Parameter 2-64
TE4 2-142, 2-165 TYPE Parameter (JOINKEYS) 2-64
TITLE 2-127 TYPE Parameter (RECORD) 2-227

Syncsort MFX Programmer’s Guide Index–13

U

UFF Format 2-33, 2-68, 2-114, 2-142, 2-157, UNIT Parameter 4-4
2-159–2-161, 2-237, 13-13 UNPAIRED Parameter (JOIN) 2-57
UNBIT 2-147 uppercase to lowercase 2-145
UNHEX 2-146 USERPASSWORD 2-128
UNINTDS Option 4-8, 5-4, 5-30 UTOL 2-146

Value-Added Products 1-7 VLFILL Parameter (OUTFIL) 2-95

Elevate PipeSort 16-2 VLTEST Option 5-3, 5-30–5-32, 14-3
Elevate PROCSort 16-1 VLTESTI Option 5-4, 5-32
Elevate ZPSaver 1-7, 16-2 VLTRAIL 2-85, 2-96
PipeSort 1-7 VLTRAIL Parameter 2-96
PROC SYNCSORT - An Accelerator for SAS™ VLTRIM Parameter (OUTFIL) 2-97
Sorting 1-7 VOL Parameter
Variable-Length Records 2-15, 2-70, 2-156, See VOLUME Parameter
2-192, 2-227, 3-9–3-11, 7-13–7-14, VOLUME Parameter 4-4
7-18–7-19, 7-34–7-35, 7-40–7-42 VSAM 4-5, 4-9, 5-26, 7-55–7-56
See also VLTEST Option See also RESET Option
Maximum Length 4-5, 12-3 RECORD Control Statement 2-227
Validity Testing 5-30 VSAM SORTOUT 5-3
VL Subparameter (JFY) 2-186–2-187, 2-189 VSAMEMT Option 5-4, 5-33
VL Subparameter (SQZ) 2-189–2-190 VTOF 2-96

WEEKDAY 2-153, 2-156 END=(conditions) 2-201

WEEKDAY Output Fields 2-156 PUSH 2-201
WHEN 2-56, 2-199–2-208, 5-32, 17-30–17-31, RECORDS=n 2-201
17-52–17-53 WHEN=INIT 2-199–2-200, 2-203, 2-207
WHEN=(conditions) 2-199–2-200, WHEN=NONE 2-199–2-200, 2-203, 2-208
2-202–2-203, 2-207 Wildcard (Pattern Constant) 2-31, 2-46,
WHEN=ANY 2-199–2-200, 2-203, 2-51–2-52
2-207–2-208 Work Areas
WHEN=GROUP 2-199–2-205, 2-207, 9-10, See SORTWKnn Statement
10-2, 17-55
Work Space 5-17, 5-25–5-27
BEGIN=(conditions) 2-200

XCTL Macro 6-3, 7-9, 7-30 10-3, 11-3

XDUP 2-21, 2-25, 4-2, 4-9, 6-2, 7-31, 7-36, 9-5, XSUM Parameter (SUM) 2-255

Y’DATEx’ 2-43–2-45 Y2C Format 2-33, 2-68, 2-71, 2-162, 2-164,

Y2B Format 2-33, 2-68, 2-71, 2-162, 2-164, 2-237, 2-240
2-237, 2-239 Y2D Format 2-33, 2-68, 2-72, 2-163–2-164,

Index–14 Syncsort MFX Programmer’s Guide

 2-237, 2-240 2-163–2-164, 2-237, 2-239, 2-244
Y2ID Format 2-142, 2-158, 2-163, 2-226, 5-7 Y2W Format 2-33, 2-44, 2-69, 2-74,
Y2IP Format 2-142, 2-158, 2-163, 2-226, 5-7 2-163–2-164, 2-237, 2-239, 2-244
Y2P Format 2-33, 2-69, 2-72, 2-163–2-164, Y2X Format 2-33, 2-44, 2-69, 2-74,
2-237, 2-240 2-163–2-164, 2-237, 2-239, 2-244
Y2S Format 2-33, 2-44, 2-69, 2-73, 2-162, Y2Y Format 2-33, 2-44, 2-69, 2-74,
2-164, 2-237, 2-241 2-163–2-164, 2-237, 2-239, 2-244
Y2T Format 2-33, 2-44, 2-69, 2-74, Y2Z Format 2-34, 2-44, 2-69, 2-71, 2-162,
2-163–2-164, 2-237, 2-239, 2-244 2-164, 2-237, 2-240
Y2U Format 2-33, 2-44, 2-69, 2-74, YD 2-152
2-163–2-164, 2-237, 2-239, 2-244 YDNS 2-152–2-153
Y2V Format 2-33, 2-44, 2-69, 2-74, Year Data, Converting 2-162

ZD Format 2-34, 2-62, 2-69–2-70, 2-164, 2-238

ZDC Output Field Format 2-143, 2-159–2-160,
13-22, 13-24
ZDF Output Field Format 2-144, 2-159–2-160,
13-22, 13-24
ZDPRINT Option 5-4, 5-33
ZIIPPCT Option 5-4, 5-33
ZPSaver 16-2
ZSPACE 1-6

Syncsort MFX Programmer’s Guide Index–15

Index–16 Syncsort MFX Programmer’s Guide