JP2828354B2 - Database management device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database management apparatus and, more particularly, to a database management apparatus for performing data search processing using a database and data update processing during data search processing in response to requests from a plurality of terminal devices. About.

[0002]

2. Description of the Related Art A database system is composed of a database and a database management device. The database management device searches for data and updates data by multitasking in response to requests from a plurality of users. I do. General data search,
Since the search is performed in real time, the search may be performed based on the latest data. However, if it is necessary to perform a search based on data at a specific time, such as a database that stores accounting data, accepting unlimited updates will cause the data to be updated one after another. Search becomes impossible.

For this reason, conventionally, when a file is updated many times, a plurality of files before updating are held as history files, and a search is performed based on the history file at a necessary time. However, in this method, since it is necessary to hold a plurality of history files before updating, a file device having a large storage capacity is required.
There is a problem.

[0004] On the other hand, when data retrieval is being performed, so-called exclusive control for eliminating a data update request is also performed. However, a user who updates data is made to wait until the data retrieval process is completed. However, there is a problem that data cannot be updated during the waiting period.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and provides a database management apparatus capable of reducing the waiting time of each user in multitasking without increasing the storage capacity of the database system. The purpose is to do.

[0006]

In order to achieve the above object, a data retrieval process using a database in response to a request from a plurality of terminal devices and a data update process during the data retrieval process according to the first aspect of the present invention. The database management device performs a storage unit that stores the updated data together with the update time, and performs a data update before the update without updating the data before the update when there is a data update request during the data search process .
A database that updates the data to correspond to the data
A first processing means for additionally stored in units of records in the storage unit stores the updated data with update time to,
And a second processing unit for performing a data search process based on data at a designated time using a database and contents stored in the storage unit when a data search request is issued. Further, according to the second aspect,
If the data before the update is not needed,
Database data is stored in the storage means
In addition to rewriting the updated data,
It is characterized by erasing memory contents.

[0007]

The database management device according to the first aspect of the present invention is provided with storage means for storing the updated data together with the update time. When there is a data update request (data addition, correction and deletion request) during the data search process, the first processing means updates the data before update without updating the data before update. Data
Is stored in the database, and the updated data is stored in the storage means together with the update time. Updated data to be stored in the storage means may be added and stored in units of records,
There is no need to provide and store a separate file. For this reason,
The storage capacity on the database system side can be reduced as compared with the conventional case of storing in file units.
The second processing means, when there is a data search request,
Using the contents stored in the database and the storage means,
A data search process based on the data at the designated time is performed for each record. Since the storage unit stores the update time, that is, the history, the data search based on the data at the time designated using the history, for example, the search based on the latest data at the time of starting the search, Can also search based on the latest data that changes every moment.

Further, according to the invention of claim 2 of the present invention ,
To, if not necessary before the update data can rewrites the updated data stored in the storage means the data of the database, it is preferable to erase the stored contents of the storage means. With this configuration, only the necessary updated data is stored in the storage unit, so that the capacity of the storage unit can be reduced.

[0009]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing the hardware of the present embodiment. A plurality of terminal devices 12 ₁ , 12 ₂ ... 12 _N are connected to a database system 10. The database system 10 includes a database and a database management device. This database system includes the index file shown in FIG.
Are stored in advance in the data file and the management program shown in FIG. The index file is composed of a plurality of records using a key (control field) and a pointer corresponding to the key as one record. The data file is composed of a plurality of records using the code, data, first pointer, correction flag, and second pointer as one record. In the case of the accounting database, the code corresponds to the account code, and the data corresponds to the amount. The first pointer indicates the address of the record. The correction flag indicates whether or not data has been corrected (corrected and deleted). When it is set, it is corrected, and when it is reset, it is not corrected. The second pointer indicates the address of the record in which the correction data is stored.

Referring to FIG. 3 as an example, the first pointers 1, 2, n, n + 1 and n + 2 represent the address of the record, and the modification flag of the record 1 is set because the correction flag of the record 1 is set. The data indicates that it has been corrected. Looking at the second pointer, since the address is n + 1, the record obtained by correcting the data of record 1 is stored in record n + 1 at address n + 1 indicated by the second pointer. Further, since the correction flag of the record n + 1 is set, the record n + 1 is also corrected, and the corrected data is stored in the record n + 2 at the address n + 2 indicated by the second pointer. Since the correction flag of the record n + 2 has been reset, the data of the record n + 2 is the latest data. That is, the data of the record 1 is corrected twice to the record n + 1 and the record n + 2. Further, since the second pointer is incremented in order of n + 1 and n + 2, it is possible to know in what order the second pointer has been modified, that is, the modification time.

Next, a processing routine on the terminal device side and a processing routine on the database system side will be described with reference to FIG. In FIG. 4, the terminal-side processing routine and the database system-side processing routine are shown side by side for easy understanding, and the steps of the terminal-side processing routine are denoted by reference characters “B”.
, And D is added to the code for the database system side steps. Hereinafter, a search based on the latest update data will be described as a real mode, and a search based on data at a specific time will be described as a balance mode.

First, a processing routine in the balance mode will be described. In step 20B, the terminal device specifies a process in the balance mode and requests a file open. On the database system side, step 22D
Accepts a file open request at step 24D
, The balance mode processing is turned on. As a result, the subsequent processing becomes the balance mode processing. In step 26D, it is determined whether or not the file is locked by use by another user. If the file is not locked, the file is opened in step 32D. On the other hand, when the file is locked, the process waits until the file lock is completed in step 28D. This file lock is determined in units of access time to one record, and a file is locked during access to one record. In the next step 30D, the file lock is released, and step 3
In 4D, the latest update number corresponding to the record of the index file, that is, the first pointer to the newest data is transmitted. In the example of FIG. 3, the record 1 is updated in order of record n + 1 and record n + 2, so that the first pointer n + 2 of record n + 2 storing the latest data, and the first pointer 2 For record n, n of the first pointer is the latest update number. On the terminal device side, step 3
The latest update number is received at 6B. Note that the terminal device is in a standby state from step 20B to step 36B, but the standby time is shorter than the conventional exclusive control because the maximum wait time is about the access time for one record. In the next step 38B, a search request in which the search number and the latest update number corresponding to the search key are assigned and the update number is specified is performed. Note that the latest update number corresponding to the search key is automatically added by the terminal device, so that the operator operating the terminal device does not need to input the update number. At step 40D on the database system side, a search request is accepted, and a search key and the latest update number are received. Then, in step 42D, a search process described later is performed and the data is transmitted. The terminal device receives the data in step 44B, and determines in step 46B whether or not the search processing has ended. If not, the process returns to step 38B to repeat the above steps. If the search process is completed, a search end request is transmitted in step 48B. As a result, the balance mode is turned off at step 50D on the database system side.

FIG. 5 shows details of the data search processing routine in step 42D. In step 60, a pointer corresponding to the search key is searched using the search key and the index file received in step 40D. In step 62, a record having the same first pointer as the pointer searched in the index file is read from the data file. In step 64, the update number of the first pointer is checked to determine whether the update number specified by the terminal device is valid. If the update number is not valid, information (EOF) indicating that the file has ended is transmitted to the terminal device in step 68. On the other hand, if the update number is valid, it is determined in step 70 whether the correction flag is set. When the correction flag is reset, the data is not corrected or deleted, so the code and data of this record are transmitted to the terminal device in step 72.

On the other hand, if it is determined in step 70 that the correction flag has been set, the second pointer, which is the updated update number, is checked, and in step 76, this update number is updated to the latest pointer designated in step 38B. Is determined to be smaller than the update number of the file, that is, whether the update number is before the file is opened or the file is unlocked.
If the updated update number is larger than the specified update number, the data before update, that is, the data of the record, is transmitted because it is data updated after the time specified in step 38B. Thereby, even if the data is updated every moment, it is possible to search for the data at the designated time and transmit it to the terminal device side.

If it is determined in step 76 that the updated update number is smaller than the specified update number, then in step 78 the history record, ie, the second
The record at the address specified by the update number of the pointer is read, and the process returns to step 70 to determine whether the correction flag is set. When the correction flag is reset, that is, when the data is not corrected, the data of the record is transmitted. When the correction flag is set, that is, when the data is corrected, the record is transmitted at step 72. Steps 74 to 78 are repeated until this data is transmitted.

FIG. 6 shows a processing routine when there is a request from another user in the balance mode. In step 80, a request from the terminal device is received, and in step 82, it is determined whether or not the request is a data addition request. At the time of a data addition request, a record including a code, data, and a first pointer is additionally stored in step 84 after the last record stored in the data file.

If a negative decision is made in step 82,
In step 86, it is determined whether the request is a correction request. In the case of a correction request, the correction flag of the record before correction, that is, the record whose data is to be corrected is set, and the address for storing the corrected data is added to the second pointer as the correction update number. At the position corresponding to the address of the second pointer, a record including the code, the corrected data, and the first pointer is additionally stored.

If the determination in step 86 is negative, it is determined in step 96 whether the request is a deletion request. At the time of the deletion request, the correction flag of the record having the data before deletion, that is, the record whose data is to be deleted is set in step 94, and the address for storing the deletion data is updated and updated in the second pointer. In step 96, a record including a code, deletion data, that is, data indicating deletion, and a first pointer whose update number is the address of this record at the address position corresponding to the second pointer is stored next to the last record. Additional storage.

As a result, during the balance mode, that is, during the data search, data update processing (data addition, data correction, data deletion) is performed. In the update processing, the updated data is updated without updating the data before the update. The record is stored by adding a new record, and data indicating that the record has been updated is stored in the record before the update. Further, since the data search is performed based on the data corresponding to the update number designated for processing, even if the data is updated from the start of the search to the end of the search, the data to be updated is not affected. Therefore, the data can be updated even during the data search, so that the waiting time of the data update user can be reduced. In addition, since the user who is performing data search can search based on the data at the start of the data search, it is not necessary to repeat the search.

FIG. 7 shows a search routine on the database system side when there is a request for the real mode. In FIG. 7, parts corresponding to those in FIG. In this routine, in steps 70 and 78, a history record is repeatedly read until a record whose correction flag is reset is read, whereby the latest data can be read in real time.

As described above, according to this embodiment, the search in the balance mode and the search in the real mode can be performed simultaneously. Therefore, it is possible to simultaneously perform two types of searches, a search using the latest data at the start of the search and a search using the latest data that changes every moment after the start of the search.

When the balance mode from all users is turned off, the database rewrites the data of the original record using the latest history record for deletion and correction and deletes the corresponding history record. 3, the data of record 1 is rewritten so that the data of record 1 becomes the data of record n + 2, and records n + 2 and n + 1 are deleted. This will erase unwanted records,
File space on the database system side can be saved. When the data is corrected a plurality of times, the latest update number (the address of the record having the latest data) is always assigned to the second pointer of the first record (the second pointer of record 1 in the example of FIG. 3). If you remember
Since the update data in the middle can be skipped and the latest record can be directly reached, the search time in the real mode can be reduced.

[0023]

According to the present invention as described above, according to the present invention, further
Corresponds to data before update without updating data before new
And store the data to the effect that the data has been updated.
Since the data is additionally stored in the storage unit in record units together with the update time, the file for saving the data before update is saved.
Therefore, even if the data update process is performed during the data search, the latest data at the start of the search and the latest data that changes every moment after the start of the search can be obtained without increasing the storage capacity. it is possible to perform data retrieval based on the data, the effect is obtained that.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of the present invention.

FIG. 2 is a diagram schematically showing an index file.

FIG. 3 is a diagram schematically showing a data file.

FIG. 4 is a flowchart showing a processing routine on the terminal device side and the database system side.

FIG. 5 is a flowchart showing details of step 42D in FIG. 4;

FIG. 6 is a flowchart showing a processing routine when another user makes a request during the balance mode.

FIG. 7 is a flowchart showing a data search routine in a real mode.

[Explanation of symbols]

10 Database system 12 _{1 to} 12 _N terminal device

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁶ , DB name) G06F 17/30

Claims (2)

(57) [Claims] 1. A database management device for performing a data search process using a database and a data update process during a data search process in response to requests from a plurality of terminal devices, and stores updated data together with an update time. Storage means and responds to data before update without updating data before update when data update request is received during data search processing
Is allowed to store the data indicative of the updated database records in the storage means the data after updated with update time
First processing means for additionally storing in units of data, and second processing for performing data search processing based on data at a designated time using data stored in the database and the storage means when a data search request is made Means, and a database management device comprising: 2. When the data before update is not required
Storing the data before updating the database in the storage means
And rewrite the updated data stored in
Erasing the storage contents of the storage means.
The database management device according to claim 1.