JPH0371262A - Data processor - Google Patents

Abstract

PURPOSE:To speed up data base arithmetic processing in a system by extracting plural fields of a record with priority based upon key information obtained from a CPU at the time of extracting a key part from the record, and processing the extracted fields as one key part. CONSTITUTION:At the time of extracting key parts from respective records of an objective file in a main storage 4 based upon an instruction outputted from the CPU 1, plural fields of one record are extracted in accordance with the priority order specified by the CPU 1 and applied to specified operation as one key part to obtain computed result. Consequently, operation based upon the multikey can be simultaneously executed, so that the data base computing processing based upon the multikey can be rapidly and efficiently executed.

Description

DETAILED DESCRIPTION OF THE INVENTION [Object of the Invention] (Industrial Application Field) The present invention relates to a data processing device having arithmetic processing functions including sorting and relational operations in a relational database.

(Prior Art) Conventionally, there is a "selection device" disclosed in Japanese Patent Application No. 62-24751 as a data processing mechanism for databases. This "selection device" is placed in the input/output channel or disk control device, and directly receives data read from a magnetic disk device used as a large-capacity data file, and selects data that satisfies conditions. It realizes an efficient data retrieval mechanism using on-the-fly processing.

However, the above-mentioned device still has the following problems. That is, in the above device configuration, when only a simple selection process is targeted, even if the target files are distributed across multiple magnetic disk devices,
A desired selection result can be obtained by adding up the information on the selection results of each magnetic disk device. However, when the target is complex operations such as sorting or relational operations in a relational database (hereinafter referred to as RDB), processing that simply adds up the information of the results of multiple magnetic disk devices as described above is required. In this case, the desired result information cannot be obtained, and each piece of target data must be subjected to arithmetic processing again, resulting in complicated processing and a concomitant reduction in processing performance, making it impossible to put it to practical use.

Furthermore, in the conventional device configuration described above, the magnetic disk device is directly connected to the selection device, and a means is adopted in which data read from the magnetic disk device is directly stored in a buffer within the selection device. Therefore, the extraction processing speed of some keys and the disk read processing speed interact with each other, and due to this, there are restrictions on file format, number of extraction keys (number of keys for multi-key), etc. There was also the problem that processing performance, processing speed, etc. were restricted.

In order to solve the above-mentioned problems, the applicant invented a "database processing mechanism" as disclosed in Japanese Patent Application No. 327542/1983.

This database processing mechanism performs predetermined arithmetic processing on a system bus on a single or multiple external memories that store files to be processed, and on databases that include sorting and relational operations in relational databases. The relational processing unit is connected to a CPU that controls the entire system, and a main memory that is accessed under the control of the CPU. A means for extracting a part of keys according to a specified key from a record of the target file stored in the first area on the main memory according to the instructions, and a means for extracting a part of keys according to the specified key, and for each of the extracted keys, position information of that key on the main memory. means for adding the key and identifier as an identifier, and performing an operation on each extraction key according to the operation instruction of the CPU using the key and identifier pair as input data; and means for sequentially extracting the identifier from the data string obtained as a result of the operation. and means for outputting to the second area on the main memory.

By using a data processing mechanism with such a configuration in a data processing system that processes databases, it is possible to easily control advanced data processing that targets complex operations such as sorting or relational operations in RDB. can be executed efficiently.

(Problem to be Solved by the Invention) In each data processing device as described above, when performing a multi-key operation, the operation is first performed using the key with the highest priority among multiple keys, and the result is , the operation is performed using the key with the second highest priority, and so on.
It is necessary to repeat the operation many times for all keys in order of priority.

The present invention has been made in view of the above-mentioned circumstances, and an object of the present invention is to provide a data processing device that can perform multi-key calculations at once, which conventionally required multiple calculations.

[Structure of the invention] (Means and effects for solving the problem) The present invention
Based on the instructions from U, when extracting a portion of keys from each record of the target file, multiple fields on one record are extracted according to the priority order specified by the CPU,
As part of one key, specified operations are performed and result information is obtained.This allows multi-key operations to be executed at once, and multi-key database operations can be executed quickly and efficiently. .

That is, the present invention provides an external storage that stores files to be processed, a processing device that performs predetermined processing on a database that includes sorting and relational operations in a relational database, and an entire system. The system has a system configuration in which a CPU that controls the CPU and a main memory that is accessed under the control of the CPU are respectively connected on a system bus, and the CPU has the processing target files stored in the external memory. and means for sending information regarding the file, calculation instructions, and specified field information to the processing unit, and the calculation processing unit includes information regarding the file, calculation commands, and specified field information. means for extracting a plurality of fields as key information from the record of the target file stored in the main memory according to the specified field information in a specified priority order, and executing an operation according to the operation command; When retrieving, multiple fields on the record are retrieved in priority order based on the key information from the CPU,
It is structured so that it is processed as part of one key, so multi-key operations can be executed at once, and database operations within the system can be executed at high speed.

Further, in the system configuration described above, the present invention provides means for converting the extracted field into data that can be processed by the arithmetic unit of the arithmetic processing device, and the arithmetic command for the converted data. and converting the field data retrieved from the main memory by DMA access according to the input format of the arithmetic unit of the arithmetic processing unit and supplying it to the arithmetic unit. It is possible to reduce the processing load on the arithmetic processing device and perform arithmetic processing such as RDB at high speed and efficiently.

(Example) The present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention. In order to simplify the explanation, here, a configuration is exemplified in which only one human output channel for controlling a magnetic disk device is provided on the system bus.

In the figure, l is the CPU that controls the entire system, and here the human output channel connected to the system bus (5-Bus) 6, the related processing unit 5, etc. are placed under control. , sorting or relational calculations in RDB, etc., the magnetic disk devices 3, 3 . ... has a control function to read the target file and write it to the first area on the main memory 4, and also has a control function to read the target file from the main memory 4 and write it to the first area on the main memory 4. file information, including information that indicates priority), and
The relational arithmetic processing unit 5 sends computation instructions for the file and instructions to a second area on the main memory 4 for storing the computation results.
It has a control function to send to. Furthermore, relational calculation processing bag b!
According to the data string (result identifier string) of the operation result written to the second area on the main memory 4 at t5, the first area on the main memory 4 is
It has processing means for rearranging the target files in the area in the order of the calculation results in the second area.

2 is an input/output channel (I
oc), which controls access to the input/output devices that are under control under the control of the CPUI, and here a plurality of magnetic disk drives 3.3... are placed under access control.

3.3. . . . are magnetic disk devices (DISI□) placed under the control of the input/output channel 2, and data groups of various file structures constituting the database are stored.

Reference numeral 4 denotes a main memory (HEM) connected to the system bus B, which includes a first area for storing files to be processed, and information regarding the files to be processed (file format, etc.) stored in this first area. A second area is provided for storing various pieces of information such as information such as information such as information such as information such as file information, record length, etc.), calculation instructions, and calculation results.

5 is a relational processing unit (
DBB), and according to information (file information) regarding the file to be processed stored in the second area of the main memory 4 and calculation instructions, the file to be processed stored in the first area of the main memory 4 is Extract multiple fields from a record in the specified priority order, convert them to a data type that conforms to the input format of the output target (operation unit), obtain the key to be subjected to operation processing, and save the first field in the main memory 4 of the record with that key. A processing function that adds location information (main memory address) to the above key as an identifier (see Figures 2 and 3), and the above operation on the key value of the same data using the key and identifier pair as input data. a processing function that executes calculations according to instructions;
It has a processing function of sequentially extracting identifiers from the data string obtained as a result of the calculation and storing them in a second area on the main memory 4, and includes a control section 51, a calculation section 52, and a buffer memory. (work buffer) 53 and other components.

Among these components, the control unit 51
According to the information (file information) about the file to be processed stored in the area and the calculation instructions, key information consisting of multiple fields is extracted from the record of the target file stored in the main memory 4, arranged in a specified priority order, and an identifier is generated. It has a control function that converts the data into data that can be processed by the calculation unit 52 and sends it to the calculation unit 52, and stores the output data received from the calculation unit 52 in a second area on the main memory 4. It has a control function to store.

That is, the control section 51 controls the main memory 4 as shown in FIG.
From the record (RFC) of the file to be processed stored in the first area above, the specified multiple fields (A, C, D) are sorted in the priority order specified above (D, C, A) to obtain key information (Ki) consisting of multiple field data, and use the head position information (main memory address) in the main memory 4 of the record having the key information as an identifier (AI). The above key (
K1), it is converted into a data type that can be processed by the calculation unit 52 (a data type that can perform binary absolute value comparison) according to the manual format (for example, binary comparison calculation format) of the calculation unit 52, and then the calculation unit In addition to having a processing means for sending data to 52,
It has a control means for storing the output data received from the calculation section 52 in a second area on the main memory 4.

The calculation unit 52 has a buffer memory (work buffer) 53
Using the above-mentioned control unit 51, the pair of key Kl and identifier AI made up of multiple field data is used as manual data, and the calculation according to the above-mentioned calculation instruction is performed on the key value of the same data. Further, the identifier AI is sequentially extracted from the data string on the buffer memory 53 obtained as a result of the calculation, and sent to the control unit 51 as output data. This output data is stored in the second area on the main memory 4 via the control unit 51.

511 to 514 are components of the control unit 51, respectively. 511 is a microprocessor (MPU) that controls the entire control unit 51, and 512 is a control memory (CPU) in which a control program for the microprocessor 511 is stored.
S). 513 is a data conversion unit 5 which converts data in the main memory 4 under the control of the microbross holder 511, which will be described later.
DMA (DirectMemor) to be transferred to 14
y Access) controller (DMAC).

514 sends the data received from the DMA controller 513 to the calculation unit 52 according to instructions from the microprocessor 511.
This is a data conversion unit that converts the data into a data type (for example, a data type that allows binary absolute value comparison) that matches the manual format of the data and sends it to the calculation unit 52.

FIG. 2 shows a microprocessor 51 in the control section 51.
1 is a flowchart showing a key extraction process flow executed under the control of FIG.

In FIG. 2, SL is a step for finding the start position of a record from the file to be processed stored in the main memory 4, and the start memory address of the record in the file to be processed in the main memory 4 is found. S2 is a step of finding the start position of the field, and the start storage address of the field in the record found in step SL is found in the order according to the file information instructed by the CPUI. S3 is DM
This is a step of issuing instructions to the A controller 513 and the data converter 514, in which data such as the position and length of the field found in step S2 is sent to the DMA controller 513, and the data type is sent to the data converter 514. . S4 is a step of detecting the last field of the target record, and it is determined whether the field found in step S2 is the last field of the plurality of fields specified by the file information from the CPUI. S5 is a step of detecting the last record of the file to be processed on the main memory 4, and it is checked whether the record found in step SL is the last record of the file to be processed placed on the TA4. to decide.

FIG. 3 shows a microprocessor 511 in the control section 51.
This is a diagram for explaining the key extraction processing means executed under the control of the taJ. The fields are A, C, and D, and the priority of the specified field is C.

A. Assuming that the identifier consisting of the head position information (main memory address) on the main memory 4 of the same record is A1, obtain a pair of key (Kl) and identifier (AI) consisting of multiple field strings, and use this key (K1). After converting the data into a data type that can be processed by the calculation unit 52 according to the input format of the calculation unit 52 (for example, binary comparison calculation format), the data is sent to the calculation unit 52.

The operation of an embodiment of the present invention will now be described with reference to FIGS. 1 to 3 above.

When the CPU receives processing targets and calculation instructions such as sorting (ascending/descending order) or relation a calculation from an input mechanism on the terminal side (not shown) via the system bus 6, the CPU receives the processing targets via the system bus B and the input/output channel 2. A file is read from the magnetic disk devices 3, 3°, . . . and written to a predetermined first area on the main memory 4. At this time, a plurality of input/output channels 2, 2 . ... are connected, the file to be processed is connected to the system bus 6 through a plurality of human output channels 2.2. The magnetic disk devices 3, 3 . . . ..., the CPUI selectively specifies the input/output channel according to the file structure of the file to be processed, issues an access instruction, and reads the file under the control of input/output channel 2. The data of the target file is stored in the main memory 4.

When the CPUI reads the target file into the first area on the taA as described above, the CPU loads the file including information about the target file (file format, block length, record length, multiple key fields and their field arrangement order, etc.). The second file on taA that stores information, calculation instructions for that file, and calculation results (result identifier string)
The area designation information and the like are sent to the relational arithmetic processing device 5, and the relational arithmetic processing device 5 is activated.

When the relational processing unit 5 receives the above information from the CPUI, it sets up the operation mode of the calculation unit 51 according to the instructions, and then performs sorting or RDB according to the instructions.
Executes relational calculation processing in .

That is, the control unit 51 in the relational arithmetic processing device 5
According to the specified contents of the file information sent from , a plurality of specified fields are extracted from each record of the file to be processed stored in the first area of the main memory 4 in the specified order, and rearranged according to the specified order. The data of multiple fields is treated as one key information (K1), and the position information (main memory address) of the key in the main memory 4 is added to each key (K1) with identifiers A1 and 17, and part of the key is Data that is a pair of Kl and identifier AI is generated. Then, the key is converted into a data type that can be processed by the calculation unit 52, and then sent to the calculation unit 52.

That is, as shown in the flowchart of FIG. 2 and the explanatory diagram of FIG. Multiple fields according to the information (A, C
, D) in the specified priority order (D, C, A) to obtain the key (Ki) of the multiple field strings (steps 81 to S
5). Furthermore, the head position information (main memory address) on the main memory 4 of the record with that key (Ki) is identified by the identifier (
A1) is added to the key (K1) (see FIG. 3), and then the data type (
data type that allows binary absolute value comparison), and the arithmetic unit 5
Send to 2.

The key extraction processing operation in the control section 51 at this time will be explained with reference to FIGS. 2 and 3.

The microprocessor 511 of the control unit 51 detects the first record storage address (record start address) from the processing target file on the main memory 4 according to the file information instructed by the CPU 1, and extracts unextracted records from that record according to the specified priority order. Field position (field start address)
is detected, the field position information, that is, the field start address is set in the DMA controller 513, and the specified data type is set in the data converter 514, and the DMA controller 513 is activated (step S4-83 in FIG. 2). ).

The DMA controller 513 is a microprocessor 51
When the field start address of the unextracted highest priority field is received from 1, the main memory 4 is accessed according to this specified address, the data of the unextracted highest priority field is read from the main memory 4, and the data is transferred to the data converter 514. do. Upon receiving the field data read and transferred from the main memory 4 under the access control of the DMA controller 513, the data conversion unit 514 converts this data into a data type (for example, binary data) that matches the input format of the calculation unit 52 and holds it. When the extraction field data for one record is completed, this data is sent to the calculation unit 52 as one key (multi-key) information. FIG. 3 shows an example of generation of key (multi-key) information by rearranging a plurality of fields at this time. Here, a plurality of fields (A, C,
D) in the specified priority order (D, C, A), obtains the key (Kl) of the multiple field string, and retrieves the starting position information (main memory address) as an identifier (^l) to the key (K1), and then input the key to the calculation unit 52 according to the input format of the calculation unit 52 (for example, binary comparison calculation format).
The data is converted into a data type that can be processed by arithmetic operations, and sent to the arithmetic unit 52.

On the other hand, when the microprocessor 511 detects the field start address of the unextracted highest priority field, it detects the start address of the next priority unextracted field. This field address search process is executed for all records of the file to be processed (Step S in Figure 2).
4. S5. St-S3).

In this way, key (multi-key) extraction processing for multiple fields and data type conversion processing are executed for each record of the processing target file. During this key (multi-key) extraction process, a flag indicating the end of the key is added in synchronization with the last data of the last designated field.

The calculation unit 52 inputs and holds the data generated by the control unit 51, and when it receives human data in which multiple fields are one key (multi-key data with matching data types), it controls the control of the control unit 51. Below, calculations are performed on each extraction key in accordance with the calculation instructions from the CPUI. Furthermore, identifiers (
AI, A2. ) is extracted to obtain output data, and the output data is sent to the control unit 51.

The control unit 51 stores the output data received from the calculation unit 52 in a pre-specified second area on the main memory 4, and completes the calculation of the identification data extracted from the target file on the main memory 4. The CPU then notifies the CPU of this fact. When the CPU receives the above calculation completion notification from the control unit 51 of the relational calculation processing unit 5, it transfers each record of the target file in the first area of the main memory 4 to the second area. Sort the calculation results (result identifier column) in the area in order, create a new file according to the specified calculation result, and save that file (
output file) information is output to the magnetic disk device 3 via the human output channel 2 and stored therein.

In the above data processing, if the size of the file to be processed exceeds the work area of the main memory 4, the file is processed for each work size of 711 and finally merged to obtain the processing result.

Further, in the above embodiment, the CPUI receives the above calculation completion notification from the control unit 51 of the related aJR calculation processing device 5, and stores each record of the target file in the first area of the main memory 4 in the abovementioned number. Sort the calculation results (result identifier column) in area 2 in order, create a new file according to the specified calculation result, and send the reconfigured file (output file) information magnetically via input/output channel 2. Disk device 3
However, by providing this processing function to the relational arithmetic processing unit 5, the CPU load can be reduced.

With the configuration of the embodiment described above, advanced data processing for complex operations such as sorting or relational operations in RDB can be efficiently executed with simple control. Also, when extracting the key part from each record of the target file based on instructions from the CPU, multiple fields on one record are extracted from the CPU in accordance with the specified priority order, and these are treated as one key part. Since specified operations are performed and result information is obtained, multi-key operations can be executed at once, and multi-key database operations can be executed quickly and efficiently.

[9! As described in detail above, according to the present invention, in a data processing device having arithmetic processing functions including sorting and relational operations in a relational database, an external storage storing a file to be processed and a sorting and an arithmetic processing unit that executes predetermined arithmetic processing on a database that includes relational arithmetic processing in a relational database; a CPU that controls the entire system;
The main memory accessed under the control of the CP is connected to the system bus, and the CP
U includes means for importing the file to be processed stored in the external storage into the main memory, and means for sending information regarding the file, calculation commands, and designated field information to the processing unit, and The arithmetic processing unit extracts multiple fields as key information from the record of the target file stored in the main memory in accordance with the information regarding the above-mentioned file, the arithmetic command, and the specified field information, and executes the arithmetic operation according to the above-mentioned arithmetic command. When extracting a key part from a record, multiple fields on the record are extracted in priority order based on key information from the CPU and processed as one key part. , multi-key operations can be executed at once, and database arithmetic processing within the system can be executed at high speed.

Further, according to the present invention, in the above system configuration, there is provided a means for converting the extracted field into data that can be processed by the calculation section of the processing unit, and a calculation command for the converted data. The field data retrieved from the main memory by DMA access is converted according to the input format of the arithmetic unit of the arithmetic processing unit, and is supplied to the arithmetic unit. It is possible to reduce the processing load on the device and perform arithmetic processing such as RDB quickly and efficiently.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention, FIG. 2 is a flowchart showing the key extraction process flow in the above embodiment, and FIG. 3 is a key extraction processing operation in the above embodiment. It is a figure for explaining. l... CPU, 2... Input/output channel (IOC),
3゜3, ... magnetic disk device (DISK), 4.
...Main memory (HEM), 5...Relational arithmetic processing unit (D
BE), 6... System bus (5-Bus), 51.
...Control unit, 52...Arithmetic unit, 53...Buffer memory (work noktufa) 511...Microprocessor (MPLI), 512...Control memory (C3)
513...DMA controller (DMAC), 5
14...Data converter, RPC...Record (A,
B, C, D. E...field), AI...identifier (main memory address), Kl...key

Claims (2)

[Claims] (1) An arithmetic processing unit that executes predetermined arithmetic processing on a database that includes an external storage that stores files to be processed and a relational operation in a relational database and sorting on the system bus; ,
A system configuration in which a CPU that controls the entire system and a main memory that is accessed under the control of the CPU are connected, and the CPU transfers files to be processed stored in the external memory to the main memory. and means for sending information regarding the file, a calculation command, and designated field information to the processing unit, and the processing unit reads the information in accordance with the information regarding the file, the calculation command, and the designated field information. A data processing device comprising means for extracting a plurality of fields as key information from a record of a target file stored in a main memory in a specified priority order and executing an operation according to the operation instruction. (2) a control unit on the system bus that executes predetermined arithmetic processing on an external storage storing files to be processed and a database including sorting and relational operations in a relational database; An arithmetic processing unit with an arithmetic unit and a C controller that controls the entire system.
A system configuration in which a PU and a main memory accessed under the control of the CPU are connected, and the CPU has:
The arithmetic processing device has means for importing a processing target file stored in the external storage into the main memory, and a means for sending information and calculation instructions regarding the file to the arithmetic processing device. A single field or multiple fields with a specified key are extracted from the record of the target file stored in the main memory according to information and calculation instructions, and the extracted fields can be processed by the calculation unit of the processing unit. 1. A data processing device comprising: means for converting the converted data into data; and means for executing an operation on the converted data in accordance with the operation instruction.