JPH09319635A - Database updating method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accelerate batch processing or update processing itself of records in a database by performing margin processing to the received update records having the same ID, and updating the records registered in the database while using the records after the margin processing. SOLUTION: A batch file managing part 20 is activated and it is checked whether the record of the same ID as a record becoming the object of update processing is defined as the object of new registration or not (exists in an update file 26 or not). When the record of the same ID exists in the update batch file 26, the batch file managing part 20 performs the marging processing for the unit of a field to the update record and the update record existent in the update batch file 26 through an update batch file access part 25, generates the new update record, deletes the existent update record and replaces the generated new update record with the deleted update record.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database updating method for executing update processing such as record registration, update and deletion in a client-server type database system connected by a network by batch processing.

[0002]

2. Description of the Related Art In recent years, client and server type database systems have been increasing with the spread of networks. In a client or server type database system connected by a network such as a LAN, when registration, update, or deletion processing for records is slow, such as when processing a large amount of data, search processing is performed online in real time. Registration, update, and deletion data update processing is performed by holding registration, update, and deletion records online in a batch file and processing the batch offline processing.

Below, such a conventional client,
The server type database system will be described with reference to the drawings. FIG. 14 shows a conventional client / server type database system.

In FIG. 14, 1 is a server machine such as WS, PC and host machine, 2 is search, registration, update,
W that has an application that requests deletion processing
Client machines such as S and PC, and the server machine 1 and the client machine 2 are connected by a network 3 such as LAN or WAN. 4 is a server machine
The database processing device 5 implemented in 1 receives a processing request from a client, receives a processing request from a client when online, and executes a search, registration, update, and deletion processing, 6 is a client machine Batch file 7 in real time online when requested to register, update, or delete records from 2
A batch file management unit 8 for executing the batch processing of the records of the batch file when offline, and an online processing unit 9 for executing the search processing of the records constituting the database 10 in real time when online. Is.

The records registered in the database 10 have the form shown in FIG. 15, for example, and one record has a record ID 11 and its data contents are "AAA", "BBB", and "CCC", respectively. It is composed of three fields 12. FIG.
Indicates one record in the batch file 7, and has a field indicating a process type 13 indicating the process of registration, update, and deletion, in addition to the record ID 11 and the field 12. The records registered in the batch file 7 are sequentially received in time series.

An outline of the operation of the client / server type database system configured as described above will be described with reference to the flow chart of FIG.

First, the application of the client machine 2 requests the database processing device 4 of the server machine 1 for record search processing, registration processing, update processing, or deletion processing (ST).
1), the database processing unit 5 accepts the request (ST2) and performs the following processing according to each processing type.

In the case of the search process, the online processing unit 9 is started (ST3), the record matching the search condition is searched from the database 10, and the search result record is returned to the client machine (ST4).

In the case of registration processing, update processing, and deletion processing, the batch file management unit 6 is started (ST5), and the registration record, update record, and deletion record are stored online in the batch file 7 as shown in FIG. Are sequentially added (ST6), and when offline, the batch processing unit 8 executes record registration processing, update processing, and deletion processing in the database 10 (ST7).

[0010]

However, in the configuration of the above-mentioned conventional client / server type database system, since batch files are created in time series,
Since update and deletion processing for the same record is executed each time during offline batch processing, there is a problem that a huge amount of time is required when a large amount of these processing occur.

As a problem of the batch processing itself,
Even if the registration to the batch file is completed, there is a problem that if the record entity in the database is not updated, it will not be reflected in the search result because it will not be a search target.

Further, due to the relationship between the data amount of records and the load on the server machine, even if the processing such as registration, update and deletion is completed in a short time so that it can be executed in real time, it is subject to batch processing and the actual update is performed. There is a problem in that the processing delays and the load of batch processing during offline increases as the number of updated records increases.

The present invention has been made in view of these problems, and records of a database are reduced by reducing the number of records subject to batch processing and the amount of record data subject to batch processing. It is an object of the present invention to provide a database updating method that realizes rapid batch processing and update processing of database records by reducing the load of batch processing itself.

[0014]

In order to achieve these objects, the database updating method of the present invention sequentially receives an update instruction and an update record for a record registered in the database, and has already received the same ID. The update record is merged and the database record is updated using the merged record.

As a result, the records in the database can be updated quickly, and the load of the batch processing itself can be reduced, so that the batch processing and the update processing of the records in the database can be speeded up.

[0016]

BEST MODE FOR CARRYING OUT THE INVENTION The invention according to claim 1 of the present invention is
The update instruction and the update record for the record registered in the database are sequentially received, the accepted update record having the same ID is merged, and the record in the database is updated using the record after the merge process. It is the one.

Further, in the invention according to claim 2, the update instruction and the update record for the record registered in the database are sequentially received, and the update record and the received update record having the same ID are merged and generated. After holding the single new update record as
The database record is updated by this new update record.

Further, in the invention according to claim 3, the update instruction and the update record for the record registered in the database are sequentially received, and a plurality of update records having the same ID that have been sequentially received are merged to perform a single process for each ID. It is configured such that the update process of the record in the database is executed after the single file.

According to the inventions of claims 1 to 3,
Batch processing, because multiple update processes for the same record can be processed at once with a single access,
It is possible to speed up the process of updating the records in the database.

According to a fourth aspect of the present invention, in the database updating method according to the first to third aspects, while the new update records generated by the merge process are sequentially stored,
The accepted update records having the same ID as the update record that is the target of the merge process are sequentially deleted. As a result, the old records that are no longer needed are deleted every time the merge process is executed, the memory that holds the updated records can be effectively used, and the batch process and the database record update process itself can be speeded up.

According to a fifth aspect of the present invention, in the database updating method according to the first to fourth aspects, when there is a registration instruction or an update instruction for the same record as the record to be finally deleted, the registration is performed. The instruction or the update instruction is invalidated. This eliminates the need to perform useless registration processing and update processing for records to be deleted later, so that batch processing and database record update processing itself can be speeded up.

According to a sixth aspect of the present invention, in the database updating method according to the first to fourth aspects, an update record before execution of update processing of a database record is stored in a registration batch file, an update batch file and a deletion batch file. It is divided and managed. As a result, the processing files are grouped and arranged in advance, so that file management is facilitated, and batch processing and database record update processing itself can be executed easily and quickly.

The invention according to claim 7 is the same as the deletion record existing in the registered batch file or the update batch file in the database updating method according to claim 6.
The record is deleted before updating the record in the database. As a result, it is possible to effectively utilize the memory that holds the update record, and to speed up the batch process and the update process of the record of the database itself.

According to the invention described in claim 8, in the database updating method according to claims 1 to 7, time management of update instruction of a record is performed, and when there are a plurality of update instructions for the same record, a new time is given. Only the update instruction of is executed. As a result, the old update process, which becomes unnecessary as a result of updating the same record again, can be deleted from the batch process target, so that the number of batch records can be reduced and the load of the batch process can be reduced.

According to a ninth aspect of the present invention, in the database updating method according to the first to seventh aspects, after the database is searched, the search result record and the update record waiting for the update process are merged. It is the one. As a result, even if the batch process for the substance of the record in the database takes a long time and the update is delayed, the result of the search executed by the client is always the latest one after the update.

The invention according to claim 10 is the method for updating a database according to any one of claims 1 to 7, wherein a time required for updating a record when receiving an update instruction of a record registered in the database is calculated, and the time is calculated. If is less than a certain value, update processing is executed in real time,
When the time is a certain value or more, the update process is executed in batch process. This allows
When the update process is completed in a short time, the update process is executed immediately, and when the update process takes a long time, the update process is executed by batch process. Will be possible.

Embodiments of the present invention will be specifically described below with reference to the drawings. FIG. 1 is an overall configuration diagram of a client / server type database system that realizes the database updating method of the present invention. In FIG.
1 is a server machine such as WS, PC or host machine, 2
Is a client machine such as WS or PC equipped with an application that requests search, registration, update, and deletion processing. The server machine 1 and the client machine 2 are
It is connected by a network 3 such as LAN or WAN.
Reference numeral 4 is a database processing device mounted on the server machine 1, 5 is a database processing unit that receives a processing request from a client, receives a processing request from a client when online, and executes search, registration, update, and deletion processing, Reference numeral 20 is a batch file management unit that creates a batch file in real time when online in response to a record registration, update, or deletion request from the client machine 2. Reference numeral 21 is a batch file that executes batch processing of records in the batch file when offline. The processing unit 22 is an online processing unit that executes, in real time, online, a search process for a record included in the database 10 and the like.
Reference numeral 23 is a registered batch file access unit for adding, updating, and deleting registration records to the registration batch file 24. Reference numeral 25 is an update batch for adding, updating, and deleting update records to the update batch file 26. The file access unit 27 is a deletion batch file access unit that adds, updates, and deletes deletion records in the deletion batch file 28.

A database updating method in the client / server type database system database processing apparatus configured as described above will be specifically described with reference to the drawings. FIG. 2 is an overall processing flow chart of the database updating method of the present invention.

First, the application of the client machine 2 sends a registration, update, and deletion record to the database processing device 4 of the server machine 1 for registration processing,
Request update processing, deletion processing, and search processing (ST20
1). The database processing unit 5 receives the registration, update, and deletion records, receives the registration processing, the update processing, the deletion processing, and the search processing request (ST202), and executes different control depending on the request content.

In the case of search processing, the database processing unit
5 starts the online processing unit 22 (ST203),
The search is executed, records matching the search conditions are acquired from the database, and the search result is returned to the application of the client machine 2 (ST204).

In the case of registration processing, the database processing section
5 starts the batch file management unit 20 (ST20
5) Then, the batch file management unit 20 adds a registration record to the registration batch file 24 by the registration batch file access unit 23 (ST206).

Next, the update process will be described. In the case of the update processing, the database processing unit 5 executes the update processing according to the processing flow chart of the update processing 1 in FIG. First, the batch file management unit 20 is activated (ST30
1), the same ID as the record that is the target of the update process
Check if the record is not subject to new registration. Specifically, the registered batch file access unit 2
3, it is checked whether or not a record having the same ID as the update record exists in the registered batch file 24 (ST30).
2).

If the registered batch file 24 has the same record, the registered batch file access unit 23 merges the update record and the existing registered record in the registered batch file 24 in field units to create a new registered record. After the generation, the existing registration record is deleted, the generated new registration record is replaced with the deleted existing registration record, and the new registration record is newly added (ST303). For example, as shown in FIG. 4, when an update record 29 for updating field 1 of a record having a record ID of 100 to "small" is received from the application of the client, the registration record 30 already registered in the registration batch file is displayed. Is stored in the registered batch file as a new registration record 31 by correcting the field 1 from “ahh” to “small”.

If there is no same record in the registered batch file 24, then the batch file management unit 20
The update batch file access unit 25 checks whether or not a record having the same ID as the update record exists in the update batch file 26 (ST304).

When the record having the same ID exists in the update batch file 26, the batch file management unit 20 causes the update batch file access unit 25 to update the update record and the existing update record in the update batch file 26 in field units. Then, the existing update record is deleted and the generated new update record is replaced with the deleted update record (ST3).
05). For example, as shown in FIG. 5, the record ID is 1
When the update record 32 for updating the field 1 of the 00 record to "small" is received from the application of the client, the field 3 of the record having the record ID 100 already registered in the update batch file 26 is "soot". The update record 33 for updating to 1 is updated to the update record 34 for updating the field 1 to “smallness” and the field 3 to “soot”.

If there is no record with the same ID in the update batch file 25, next, the batch file management unit 20
The delete batch file access unit 27 deletes the record having the same ID as the update record from the delete batch file 28.
Then, it is checked whether or not (ST306). Same ID
If the record of is in the deletion batch file 28,
Since the update record is a record that does not need to be updated,
The batch file management unit 20 discards the received update record (ST307).

If there is no record with the same ID in the deleted batch file 28, the batch file management unit 20 will
The update batch file access unit 25 adds the received update record to the update batch file 26 (ST308).

In this way, the same ID as the update record
Check whether the record of exists in the registered batch file and the deleted batch file. If the same record already exists in the registered batch file or the update batch file, both records are merged field by field By executing the process to update the records in the batch file every time the update record is accepted, the registration process and update process for the same record can always be managed and processed based on a single record. It is possible to dramatically reduce the number of actual records to be processed. Also, if it is confirmed that the deleted batch file already holds the record with the same ID, all the registered records and update records with the same ID are deleted, so the number of batch records can be significantly reduced. Become.

Next, the deletion process will be described. In the case of the deletion process, the database processing unit executes the deletion process according to the flow chart of the deletion process 1 in FIG. First, the database processing unit 5 activates the batch file management unit 20 (ST601).

The batch file management unit 20 uses the same ID as the deleted record by the registered batch file access unit 23.
Is checked in the registered batch file 24 (ST602). If there is, the batch file management unit 20 deletes the record from the registered batch file 24 (ST603).

If not, the batch file management unit 20 next checks whether or not the update batch file 26 has a record having the same ID as the deleted record (ST60).
4). If there is, the batch file management unit 20 deletes the record from the update batch file 26 (S
T605).

If not, then the batch file management section 20 checks whether the same record as the deleted record exists in the deleted batch file 119 (ST60).
6). In some cases, the deleted records are duplicated and do not need to be re-registered. Therefore, the batch file management unit 20 discards the received deleted record (ST.
607).

When there is no record with the same ID in the deletion batch file 28, the batch file management unit 20 first adds the received deletion record to the deletion batch file 28 (ST608).

In this way, the same ID as the deletion record existing in the registered batch file or the update batch file
Since the registration process and the update process itself are not required for the record No., the number of batch records can be reduced by deleting all of them. For example, FIG.
When the update processing is received in time series as described above, the batch processing in the conventional database processing apparatus requires processing for all 8 records in total, but according to the present invention, every time the update processing is received, As a result of retrieving records with the same ID and executing merge processing and performing batch file aggregation processing, the number of records subject to batch processing is shown in FIG.
As shown in, there are a total of 4 records.

By the above processing, batch files waiting for registration, update, and deletion in the database 10 are aggregated into a single record for each record ID. Therefore, in the present invention, the batch files are retrieved from the client machine 2. When a request is made, it will be reflected in the search results.

This processing is executed as a part of the function of the online processing unit 22. This will be specifically described with reference to the search processing flow chart of FIG. First, when the search processing request is received from the client machine 2 application, the database processing unit 5 activates the online processing unit 22 (ST901), executes the search, and acquires and retains the record that matches the search condition from the database 10. Yes (ST9
02).

Next, the online processing unit 22 extracts one record from the set of records as the search result (ST
904), start the batch file management unit 20 (ST
905), it is checked whether or not each batch file has a record with the same ID.

The batch file management section 20 first causes the deletion batch file access section 27 to check whether the deletion batch file 28 has a record with the same ID (ST906). If there is, this is notified to the online processing unit 22 (ST907), and the online processing unit 22 deletes the corresponding record from the search result record (ST908).

If there is no record with the same ID in the deleted batch file 28, next, the batch file management unit 20
The update batch file access unit 25 checks whether or not there is a record with the same ID in the update batch file 26 (ST909).

When the update batch file 26 has a record with the same ID, the update batch file access unit 25
Checks whether the update record of the update batch file 26 matches the search condition (ST910), and if it matches, notifies the online processing unit 22 that the record with the same ID exists in the update batch file 26. (ST
911), the online processing unit 22 executes the merge process of the search result record and the update record of the update batch file 26, and updates the search result record with the newly generated new update record (ST912). If they do not match, the online processing unit 22 is notified and the search result record is discarded (ST913). This is because the actual record in the database 10 has not been updated because the batch process has not yet been executed, but a record update request from any one of the clients 2 has already been accepted, and the current search result record is originally a search condition. Is a record that does not match.

If there is no record with the same ID in the update batch file 26, similar processing is executed for the next search result record (ST914).

In this way, the processing from ST904 to ST914 is repeatedly executed until there are no search result records retrieved from the database (ST903).

After exiting the processing loop of the deletion batch file 28 and the update batch file 26, the online processing unit 22 executes the search for the registration record of the registration batch file 24 by the registration batch file access unit 23. A registration record that matches the search condition is acquired (ST915), and the online processing unit 22 adds the corresponding record to the search result record (ST916).

After the above search processing is executed, the database processing unit 5 sends the search result record to the application of the client.

By the above-mentioned processing, even before the actual record of the database 10 is updated, the result is the search target. Therefore, while adopting the batch processing for the update of the record,
The latest search result can be provided to the client, and the real-time property of the search can be secured.

Next, a second embodiment of the present invention will be described. In the present embodiment, the batch processing unit 21 further
The execution timing of the update processing of each batch file for the database 10 is adjusted considering the performance of the entire system. When the update processing of the database 10 is completed in a short time, it is executed online, and when it takes time, It is supposed to be executed by batch processing. This will be specifically described with reference to the process flow charts of FIGS. 10 and 11.

This processing is also executed as a part of the function of the online processing unit 22. First, client machine 2
Application requests registration processing, update processing, deletion processing, and search processing to the database processing apparatus 4 (ST1001), the database processing unit 5 accepts this (ST1002), and executes this if search processing. Yes (ST1003).

In the case of registration processing, the database processing unit 5 acquires the CPU load status and the number of records in the database, calculates the registration processing time under the current load from these values, and determines whether processing is possible within the specified time. Check (ST100
4). At that time, for example, a specified time for registration processing, a real-time processing time without load [per unit record length],
The processing time increase ratio [vs CPU load] and the processing time increase ratio [vs unit DB record number] are set in advance as in the processing time table of FIG. 12, and the calculation formulas thereof are set as shown in FIG.

When the prescribed time is required, the database processing unit 5 causes the online processing unit 22 to execute the registration process on the database 10 in real time (ST10).
05) If the specified time is exceeded, registration processing is performed in the registration batch file as in the normal case (ST1006).
To ST1007).

For example, the registration record length is 200 bytes,
The CPU load of the server machine acquired from the OS is 30%,
Assuming that the number of database records is 100, the real-time registration processing time per record under the current load is 6.8 seconds (in the setting of FIG. 12, the calculation formula of FIG. 13 is used). If the specified time for the preset registration process is 7 seconds, the registration process can be performed in the specified time, and thus the registration process is registered in the database 10 in real time.

In the case of the update processing, the database processing unit 810 acquires the CPU load status and the number of records in the database, calculates the update processing time under the current load from these values, and performs the real-time update processing under the current load. It is checked whether processing can be performed within a specified time (ST1010). When the prescribed time is required, the database processing unit 5 causes the batch file management unit 20 to check whether the same record exists in the registered, updated, and deleted batch files (ST10).
11). If it exists, updating the record of the database 10 online causes a mismatch with the record of each batch file. Therefore, the update process for each batch file including the merge process already described is executed (ST1012). If it does not exist, update processing is executed in real time (ST1013).

When the specified time is exceeded, the update process for each batch file is executed in the same manner as the normal update (ST1014).

Also in the case of the deletion processing, similarly, the deletion processing time is calculated and it is checked whether the processing can be performed within the specified time (ST.
1015). If the specified time is required, the batch file management unit 20 checks whether or not the same record exists in the registration, update, and deletion batch files (ST1016), and if there is, delete processing for each batch file is executed (ST1017). If it does not exist, deletion processing is executed in real time (ST1018). If the specified time is exceeded, delete processing is executed for each batch file (ST1019).

As described above, when the data amount of records and the load on the server machine are small and the registration, update, and deletion processing in the database in real time can be completed in a short time, the processing for the database record should be executed in real time. This makes it possible to reduce the load of batch processing when offline.

[0065]

As described above, according to the present invention, a plurality of update processes for the same record can be processed at a time by a single access. Therefore, batch process and database record update process itself can be speeded up.

Also, the old record that is no longer needed is deleted every time the merge process is executed, the memory for holding the update record can be effectively used, and the batch process and the update process of the record of the database itself can be speeded up.

Further, since it is not necessary to execute useless registration processing and update processing for records to be deleted later, batch processing and update processing of database records can be speeded up.

Further, since the processing files are grouped and arranged in advance, file management is facilitated, and batch processing and database record update processing itself can be simply and quickly executed.

Further, the memory for holding the update record can be effectively used, and the batch process and the update process of the record of the database itself can be speeded up.

Further, since the old update process, which becomes unnecessary as a result of updating the same record again, can be deleted from the batch process target, the number of batch records can be reduced and the load of the batch process can be reduced.

Further, even if the batch processing for the substance of the record of the database takes time and the update is delayed, the result of the search executed by the client is always the latest after the update.

When the update process is completed in a short time, the update process is executed immediately, and when the update process takes time, the update process is executed by the batch process. Optimal operation is possible as a whole.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram of a client / server type database system that realizes a database updating method of the present invention.

FIG. 2 is an overall processing flow chart of the database updating method of the present invention.

FIG. 3 is an update processing flow chart of the database updating method of the present invention.

FIG. 4 is a diagram showing records in an update batch file of the present invention.

FIG. 5 is a diagram showing records in an update batch file of the present invention.

FIG. 6 is a deletion processing flow chart of the database updating method of the present invention.

FIG. 7 is a diagram showing a time series update request record of the present invention.

FIG. 8 is a diagram showing records in each batch file of the present invention.

FIG. 9 is a search processing flowchart of the database updating method of the present invention.

FIG. 10 is a process flow chart of a database updating method according to a second embodiment of the present invention.

FIG. 11 is a flowchart of the processing (continuation) of the database updating method according to the second embodiment of the present invention.

FIG. 12 is a diagram showing a processing time table used in the database updating method according to the second embodiment of the present invention.

FIG. 13 is a diagram showing a processing time calculation method used in the database updating method according to the second embodiment of the present invention.

FIG. 14 is an overall configuration diagram of a client / server database system that realizes a conventional database updating method.

FIG. 15 is a diagram showing records in a conventional database.

FIG. 16 is an overall processing flowchart of a conventional database updating method.

FIG. 17 is a diagram showing records in a conventional batch file.

[Explanation of symbols]

 10 database 20 batch file management unit 21 batch processing unit 22 online processing unit 24 registered batch file 26 update batch file 28 deletion batch file

Claims (10)

[Claims] 1. An update ID and an update record for records registered in a database are sequentially received, and the same ID is received.
A method for updating a database, wherein a merge process is performed on a received update record having the following, and a record in the database is updated using the record after the merge process. 2. A single new update record generated by merging an update instruction for a record registered in a database and an update record in sequence and merging the update record and an accepted update record having the same ID. A database update method characterized by updating the record of the database with this new update record after holding it as an update record. 3. An update instruction for a record registered in a database and an update record are sequentially received, and a plurality of update records having the same ID that are sequentially received are merged to form a single file for each ID. A database updating method characterized by executing update processing of a record of the database. 4. The new update record generated by the merge process is sequentially stored, while the accepted update record having the same ID as the update record targeted for the merge process is sequentially erased. 1 to claim 3
Database update method described. 5. When the registration instruction or update instruction for the same record as the record to be finally deleted is given, the registration instruction or update instruction is invalidated. Database update method described. 6. An update record before execution of update processing of a record in a database is divided into a registration processing batch file, an update processing batch file, and a deletion processing batch file for management. Database update method described. 7. The database updating method according to claim 6, wherein the record having the same ID as the deleted record existing in the registered batch file or the update batch file is erased before the record update of the database. 8. The record update instruction is time-managed, and when there are a plurality of update instructions for the same record, only the update instruction of a new time is executed. Database update method. 9. The database update method according to claim 1, wherein after the search is performed on the database, the search result record and the update record waiting for the update process are merged. 10. The time required to update a record when receiving an update instruction for a record registered in the database is calculated. If the time is less than a certain value, the update process is executed in real time and the time is updated. 10. The database updating method according to claim 1, wherein the updating process is executed in a batch process when is greater than or equal to a certain value.