JP2002297423A - Write delay database management method, device, program and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a database a management system, for managing a relational database, that can enhance update performance during the system operation while maintaining the performance capable of high speed response to sophisticated search request. SOLUTION: A data-for-search storage means 4 processes a search operation at a high speed and processes a modification operation at a low speed. Data-for- insertion storage means 1 (5), 2 (6) and data-for-deletion storage means 1 (7), 2 (8) process the search operation at a low speed and process the modification operation at a high speed. A data transfer means 9, for reflecting the data processed with the data storage means 4 to 8, transfers the data, to the data storage means 4, from the data storage means 5 to 8. A database operation request processing means 3 distributes the inputted operation requests for the database to the respective data storage means 4 to 8 to execute the requests. A transaction processing means 2 ensures, between the data transfer means 9 and the database operation request processing means 3, the data consistency.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a database management method, system, program, and recording medium, and more particularly, to a database system in which updating and retrieval frequently occur, such as a document management system. The present invention relates to a database management method, system, program, and recording medium for performing database update and search processing.

[0002]

2. Description of the Related Art A relational database handles data by expressing it in a table. A table is a set of tuples, and a tuple is a sequence of attribute values. The entity of the table is stored in the file. Operations on the database can be divided into search operations, insert operations, update operations, and delete operations. The search operation is an operation of giving a condition regarding an attribute value as a search condition and obtaining a set of tuples that match the condition. The insert operation is an operation for inserting a new tuple having a given attribute value into a table. The update operation is an operation for changing the attribute value of the tuple selected from the table to a new value. The delete operation is an operation for deleting the tuple selected from the table. The insert (Insert) operation, the update (Update) operation, and the delete (Delete) operation are collectively referred to herein as a change operation.

One of the important performances in a system using a database represented by a relational database is a response time to a search operation. A method used to reduce the response time to a search operation is the introduction of an index file. The index file can be generalized as one that converts one or a plurality of attribute values into a specific structure so that conditions relating to those attribute values can be quickly evaluated.

[0004] On the other hand, in the change operation, the time required for updating the index file causes the performance to deteriorate. Until now, performance measured by response time to a change operation has not been valued. That is, the frequency of requesting change operations is lower than that of search operations, and as a usage form, usually only search operations are performed, and a large amount of change operations can be performed with the system stopped at night etc. Depending on.

However, when real-time properties are required as in an online system, response time to a change operation is regarded as important. To deal with such a problem, for example, in a “database management system” disclosed in Japanese Patent Application Laid-Open No. H10-143412, a write to a database is temporarily stored in a non-volatile memory before being reflected on a magnetic disk, and For reference, a non-volatile memory is used as a disk cache instead of a magnetic disk. However, there is a problem that only data having a simple structure can be stored in the disk cache, and a high-performance index file cannot be used.

When a plurality of users simultaneously use a database system, a search operation and a change operation are requested asynchronously. At this time, transaction processing is used to maintain data consistency. The transaction processing is described in detail in [1] "Introduction to Transaction Processing System" by Philip A. Bernstein and Eric Newcomer Nikkei BP.

While complete isolation of transactions ensures data consistency at any point in time, reduced concurrency can significantly reduce overall throughput. To avoid this problem, the concept of isolation level is used. For the isolation level, refer to [2] “A Criticu
e of ANSI SQL Isolation L
evels "HalBerenson, Phili
pA Bernstein, Jim Gray,
Jim Melton, Elizabeth J.
O'Neil, Patrick E.L. O, Neil
Proc. ACM SIGMOD Conf. (J
un. 1995) p. 1-10 ".

[0008] As described above, as the amount of data handled by the database becomes large, it is required that the system be continuously operated without stopping. ,
While maintaining the ability to respond quickly to advanced search requests,
We proposed a database management system and system that can improve the update performance during system operation.

However, Japanese Patent Application No. 2000-25247.
In the database management system and system in the specification of Japanese Patent No. 2, change processing (insertion, update, deletion) is processed by one data holding unit. Therefore, an attribute value called a deletion flag is prepared to identify a deleted tuple. When obtaining the search results, it is necessary to check the deletion flags of all the tuples of the obtained result set and remove them from the results.

[0010]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and further improves update performance during system operation while maintaining performance capable of responding to advanced search requests at high speed. It is an object of the present invention to provide a database management method, system, program, and recording medium that can be made to operate.

[0011]

In the database management method, system, program, and recording medium according to the present invention, another data holding means for deletion is used for deletion processing. The same mechanism as the insertion data holding means (change data holding means in the invention described in Japanese Patent Application No. 2000-252472) can be used for this data holding means. When the search result is obtained, only the set operation for the set of tuple IDs obtained from the data holding means for search, insertion, and deletion is sufficient, and it is not necessary to refer to the attribute value of the tuple, and the processing can be performed efficiently.

According to a first aspect of the present invention, there is provided a database management apparatus for managing a database, comprising: a search data holding unit that performs high-speed data search operations but low-speed data change operations; Is a high-speed first and second insertion data holding unit; a data change operation is a high-speed first and second deletion data holding unit; and the first and second insertion data are In order to reflect the data processed by the holding unit and the first and second data for deletion, the first and second data for insertion and the first and second data for deletion are reflected. A data transfer unit for transferring data from a holding unit to the search data holding unit; and a first and second insertion request for an operation request for the database according to the content of the operation request. A database operation request processing means for executing the data transfer means separately to each of the first and second data storage means for deletion and the data storage means for search; the data transfer means and the database operation request processing means; And transaction processing means for guaranteeing data consistency between the two.

According to a second aspect of the present invention, in a database management apparatus for managing a database, a database operation request input means for inputting a database operation request from a plurality of users, and a high-speed data retrieval operation process, The process of the change operation is a low-speed search data holding unit, the process of the data change operation is a high-speed first and second insertion data hold unit, and the process of the data change operation is the first and second high-speed process. , And the first and second data storage means for reflecting the data processed by the first and second data storage means for insertion and the first and second data storage means for deletion. 2 data transfer means for transferring data from the data holding means for insertion to the data holding means for search from the first and second data holding means for deletion; An operation request for the database input to the database operation request input means is transmitted to the first and second insertion data holding means and the first and second deletion data holding means in accordance with the content of the operation request. And a database operation request processing unit that executes the operation by distributing the data to each of the search data holding units, and a transaction processing unit that guarantees data consistency between the data transfer unit and the database operation request processing unit. It is characterized by the following.

According to a third aspect of the present invention, in the first or the second aspect of the present invention, when the data insertion request is received while the data of the first insertion data holding unit is being inserted into the search data holding unit. And inserting the data requested to be inserted into the second data storage unit for insertion.

According to a fourth aspect of the present invention, in the first or the second aspect of the present invention, when there is a data deletion request while deleting the data of the first data holding means for deletion from the data holding means for search. And inserting the data requested to be deleted into the second data-for-deletion holding means.

According to a fifth aspect of the present invention, in the first aspect of the present invention, the search data holding means comprises:
Any one of the first insertion data holding unit, the second insertion data holding unit, the first deletion data holding unit, and the second deletion data holding unit.
Alternatively, the plurality is characterized in that the database management system itself can be configured to operate independently.

According to a sixth aspect of the present invention, in the first aspect of the present invention, the first insertion data holding unit, the second insertion data holding unit, and the first deletion data are stored. One or more of the holding unit and the second data holding unit for deletion are provided on a memory.

According to a seventh aspect of the present invention, in accordance with any one of the first to fourth aspects of the present invention, the search data holding means comprises:
The first data storage unit for insertion, the second data storage unit for insertion, the first data storage unit for deletion, and the second data storage unit for deletion are configured according to the characteristics of attribute values. Is selected.

According to an eighth aspect of the present invention, in accordance with any one of the first to fourth aspects of the present invention, the data transfer unit holds the designated transfer time information, so that the data transfer unit holds the search data. The data change operation to the means can be delayed until the time specified in the transfer specified time information.

According to a ninth aspect of the present invention, in the first aspect of the present invention, the data transfer unit is configured such that a change operation request to the search data holding unit to be executed by the data transfer unit is constant. It is characterized in that a change operation is executed every time the number reaches the number.

According to a tenth aspect of the present invention, in any one of the first to fourth aspects, the change operation for a plurality of processing unit data in the search data holding means is executed as a batch process. Things.

According to the eleventh aspect of the present invention, there is provided a data holding means for search which performs a high-speed data search operation but has a low-speed data change operation, and a first and a second high-speed data change operation which perform a high-speed data change operation. A database management method for processing and managing a database using data holding means for insertion and processing for changing data by using first and second data holding means for high-speed deletion. From the first and second data storage means for insertion and the first and second data storage means for deletion in order to input a database operation request from the server and reflect the processed data. Transfer the data to the data holding means, the operation request for the input database,
In accordance with the content of the operation request, the first and second data storage means for insertion, the first and second data storage means for deletion, and the data storage means for search are respectively distributed and processed. It is characterized by ensuring data consistency between data transfer and processing of the database operation request.

According to a twelfth aspect of the present invention, in the invention of the eleventh aspect, when a request for data insertion is made while data of the first data storage means for insertion is being inserted into the data storage means for search, It is characterized in that the data requested to be inserted is inserted into the second insertion data holding means.

According to a thirteenth aspect of the present invention, in the invention of the eleventh aspect, when there is a data deletion request while deleting the data of the first deletion data holding means from the search data holding means, It is characterized in that the data requested to be deleted is inserted into the second data storage means for deletion.

The invention of claim 14 is the invention of claims 11 to 13
In any one of the inventions described above, the search data holding unit, the first insertion data holding unit, the second insertion data holding unit, the first deletion data holding unit, and the second deletion Any one or more of the data holding means can independently configure the database management system so as to operate independently.

The invention of claim 15 is the invention of claims 11 to 13
Any one of the first insertion data holding unit, the second insertion data holding unit, the first deletion data holding unit, and the second deletion data holding unit. One or more of them are on a memory.

The invention of claim 16 is the invention of claims 11 to 13
In any one of the inventions described above, the search data holding unit, the first insertion data holding unit, the second insertion data holding unit, the first deletion data holding unit, and the second deletion The destination data holding means is characterized in that its form is selected according to the characteristics of the attribute value.

The invention of claim 17 is the invention of claims 11 to 13
In any one of the inventions described above, by holding designated time information for designating the time of the data transfer, the data change operation to the search data holding means at the time of the data transfer is performed by the designated time information. Is delayed until the time specified in the.

The invention according to claim 18 is the invention according to claims 11 to 13
In any one of the inventions described above, the data transfer is characterized in that a change operation is executed each time a change operation request to the search data holding unit to be transferred reaches a certain number.

The invention of claim 19 is the invention of claims 11 to 13
In any one of the inventions described above, the change operation for a plurality of processing unit data in the search data holding unit is executed as a batch process.

According to a twentieth aspect of the present invention, there is provided a database management method according to any one of the first to tenth aspects for functioning as each means of the database management apparatus according to the first aspect, or the database management method according to any one of the eleventh to nineteenth aspects. This is a program to be executed.

According to a twenty-first aspect of the present invention, there is provided a computer-readable recording medium storing the program according to the twentieth aspect.

[0033]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (System Configuration) FIG. 1 is a block diagram for explaining the functions of a database management system according to the present invention. Hereinafter, a relational database will be described as a representative, but the present invention is also applicable to a normal database. An operation request for the database is input from the database operation request input means 1,
The database operation request processing means 3 executes the operation. The search data holding unit 4 processes a database search operation. In addition, the first insertion data holding unit 5 (referred to as insertion data holding unit 1; hereinafter similarly expressed).
The data holding means for insertion 2 (6) processes data insertion operation, deletion operation and update operation of the database, and the data holding means for deletion 1 (7) and the data holding means for deletion 2 (8) store data of the database. Handles insert and delete operations. The data holding units 4 to 8 described here are, for example, in a relational database, an index file and a unit for performing processing using the index file. Means for performing various processes in accordance with the association. The data transfer means 9 comprises insertion data holding means 5, 6 and deletion data holding means 7, 8.
Processing unit data (a tuple in a relational database) is read from the
Insert, update, and delete tuples for. The transaction (exclusive control) processing means 2 performs transaction control in the system. Each of these elements will be described in further detail below.

(Database Operation Request Input Means) FIG.
FIG. 1 is a diagram showing an example of a system configuration embodying the present invention.
It is realized as. A server host 10 holding a database has a CPU 11, a memory 12 having a program area 12a and a data area 12b, and a hard disk 13 connected by a bus 14. A plurality of input terminals 21 are connected to the server host by a LAN 22. It is possible.

The user inputs a database operation request represented by a character string in the form of, for example, an SQL statement from the input terminal 21. The database operation request is sent to the server host 10 via the LAN 22 for processing. The result of the processing is sent again to the input terminal 21 via the LAN 22 and displayed on the display of the input terminal 21 to notify the user.

(Database Operation Request Processing Unit) FIGS. 3 to 5 are flowcharts showing an example of a series of processing procedures of the database operation request processing unit. The processing contents differ depending on the type of the database operation request. In this example, each tuple in the table is given a uniquely determined ID (tuple ID), and all tuples are identified by the tuple ID.

When the database operation request processing means 3 receives the database operation request processing (step S1), it judges whether or not the operation is an insertion operation (step S2). In the case of the insertion operation, after acquiring a tuple ID for a new tuple, (Step S3) It is determined whether the insertion data holding unit 1 (5) is reflecting the data in the search data holding unit 4, that is, whether it is in use (Step S4). If the data is being reflected, a tuple insertion operation is performed on the insertion data holding unit 2 (6) (step S5), and the result is returned (step S7). If the data is not being reflected, the insertion operation is performed on the insertion data holding unit 1 (5) (step S6), and the result is returned (step S7).

If it is not an insertion operation in step S2, the data holding means for search 4, the data holding means 1 for insertion
(5), data holding means for insertion 2 (6), data holding means for deletion 1 (7), data holding means for deletion 2
A search operation is performed for each of (8) (steps S8 to S12 in FIG. 4), and each search result set Rr, Ri
1, Ri2, Rd1, Rd2, a final result set R (R
= Rr + Ri1 + Ri2-Rd1-Rd2;
Represents a logical sum, and − represents a logical difference. ) Is made as a search result (step S13).

Next, if it is a search operation in step S14, the result is returned (step S7), and if it is not a search operation, it is further determined whether or not the search result set Ri1 exists (step S15). If not in Ri1, search result set Ri
2 is determined (step S17). Ri1, R
If it exists in i2, a delete operation is performed on the data storage means for insertion 1 (5) and the data storage means for insertion 2 (6), respectively (steps S16 and S18).
In this way, in the deletion operation, when the tuple selected in the same procedure as the search operation is in Ri1, the deletion processing is executed on the data storage unit for insertion 1 (5), and when the tuple is in Ri2, A deletion process is performed on the data storage unit for insertion 2 (6).

When neither Ri1 nor Ri2 exists, ie, R
If it is in r (if it does not exist in Ri2 in step S17), the process proceeds to step S19 in FIG. In other words, it is determined whether the data storage unit for deletion 1 (7) is in use (step S19), and the data storage unit for deletion 1
If (7) is reflecting the data in the search data holding unit 4, the tuple is inserted into the deletion data holding unit 2 (8) (step S20). If it is not being reflected, a tuple insertion operation is performed on the data storage unit for deletion 1 (7) (step S21).

Thereafter, it is determined whether the operation is an update operation (step S22). If the operation is an update operation, the tuple is updated to a new value (step S23). It is determined whether the data storage unit for insertion 1 (5) is in use (step S24). If the data is being used, a tuple is inserted into the data storage unit for insertion 2 (6) (step S25). If not, a tuple insertion operation is executed in the insertion data holding unit 1 (5) (step S26), and the process proceeds to step S7. Thus, in the update operation, the tuple selected in the same procedure as the search operation is subjected to the delete operation by the same means as the delete operation, and the tuple changed to the new value is subjected to the same operation as the insert operation. The update operation is executed by executing the insert operation by means of (1).

(Search Data Holding Means) The search data holding means 4 processes the search operation executed by the database operation request processing means 3. Further, it processes a change operation performed by the data transfer means 9. The search data holding unit 4 can execute the search operation at high speed, but the change operation is relatively slow. For example, it is conceivable to use a full-text search server that can perform a full-text search.

(Insertion Data Holding Means 1 and 2) The insertion data holding means 1 (5) and 2 (6) process the insertion operation, deletion operation and update operation executed by the database operation request processing means 3. Further, it processes a delete operation performed by the data transfer means 9. Data holding means for insertion 1
(5) and 2 (6) can execute the change operation at high speed.
For example, a normal file managed by the OS may be used.

When the insertion data holding means 1 (5) and 2 (6) cannot execute the search operation as in a normal file, the held tuples are sequentially returned, and the database operation request processing means 3 performs the search. Evaluate the condition (called an exhaustive search). Insertion data holding means 1 (5) and 2
The number of tuples held in (6) is the search data holding means 4
Is very small compared to the number of tuples held by, so that a 100% search does not affect the response time.

(Data Holders 1 and 2 for Deletion) The data holders 1 (7) and 2 (8) for deletion process the insertion operation executed by the database operation request processing means 3. Further, it processes a delete operation performed by the data transfer means 9. The data holding means for deletion 1 (7) and 2 (8) can execute the change operation at high speed. For example, a normal file managed by the OS may be used.

When the data holding means for deletion 1 (7) and 2 (8) cannot execute the search operation as in the case of a normal file, the held tuples are sequentially returned and the database operation request processing means 3 searches the data. Evaluate the condition (called an exhaustive search). Data holding means for deletion 1 (7) and 2
The number of tuples held by (8) is the search data holding unit 4
Is very small compared to the number of tuples held by, so that a 100% search does not affect the response time.

Any one or more of the data holding means 4 to 8 may be capable of independently configuring a database management system so as to operate independently.
Any of such data holding means 4 to 8 may be applied to the database management system of the present invention. That is, the response to the update request can be improved without making a special change to the database management system for a search that is fast but relatively slow in the existing database management system. . Conversely, if the update response of the existing database management system is fast, the response to the search request can be improved while utilizing the update response performance by combining it with a high-speed data holding means.

Also, the data holding means for insertion 1 (5),
2 (6), data holding means for deletion 1 (7), 2 (8)
May be stored in the memory, and when there is a large amount of memory, the data holding means for insertion 1 (5), 2 (6),
By using the data holding means for deletion 1 (7) and 2 (8) on the memory, remarkable improvement in update performance can be realized.

The data holding means 4 for search, the data holding means 1 (5), 2 (6) for insertion and the data holding means 1 (7), 2 (8) for deletion correspond to the characteristics of attribute values. It is preferable to select the mode. For example, for the attribute value having a small data amount, the data holding means for insertion 1 (5) and 2 (6) and the data holding means for deletion 1 (7) are stored in the memory as described above. , 2 (8) can improve the performance.

(Data Transfer Means) FIG. 6 is a flowchart for explaining an example of the processing procedure of the data transfer means. The data transfer means 9 is the data holding means 1 for deletion.
Tuples are sequentially read from (7) (step S31),
According to the value, the tuple is deleted from the search data holding unit 4 (step S32), and the tuple that has been reflected is deleted from the deletion data holding unit 1 (7) (step S3).
3). Similarly, tuples are sequentially read from the data storage unit for deletion 2 (8) (step S34), and are deleted from the data storage unit for search 4 according to the values (step S34).
35) The tuple that has been reflected is deleted from the data storage means 2 (8) for deletion (step S36). Next, tuples are sequentially read from the insertion data holding unit 1 (5) (step S37), and the tuples are inserted into the search data holding unit 4 (step S38). 1 (5) (step S39). Then, similarly, tuples are sequentially read from the insertion data holding unit 2 (6) (Step S).
40), the tuple is inserted into the data storage unit for search 4 (step S41), and the tuple that has been reflected is deleted from the data storage unit for insertion 2 (6) (step S4).
2).

By designating the transfer time information to the data transfer means 9, the operation of changing the data to the search data holding means 4 by the data transfer means 9 can be delayed until the specified time. Thus, when the time period during which the operation rate of the system becomes low is determined, the throughput of the entire system can be increased by concentrating data reflection on the search data holding unit 4 in that time period.

Further, the data transfer means 9 executes the change operation every time a change operation request to the search data holding means 4 reaches a certain number, so that the update request to the search data holding means 4 is issued. This can reduce the frequency of occurrence, thereby reducing the frequency of the need for exclusive control and increasing the throughput of the entire system.

Further, by making the change operation for a plurality of tuples in the search data holding unit 4 executed as a batch process, if the operation can be executed at high speed, the operation of the data transfer unit 9 can be executed at high speed. As a result, the throughput of the entire system can be increased.

(Transaction processing means) The transaction processing means 2 performs exclusive control for controlling the execution order of the asynchronously requested search operation and change operation so as to maintain data consistency, and when the change operation is stopped. Performs logging to record information for returning to the state before the transaction started. The exclusive control is used not only for exclusive control between database operations requested by a plurality of users but also for exclusive control between the database operation request processing means 3 and the data transfer means 9.

Various isolation levels are realized by using the multi-granular lock described in the aforementioned reference [1]. Locks are applied to tables, tuples, and files. The file in the present invention can be regarded as being composed of search data holding means, insertion data holding means 1 and 2, and deletion data holding means 1 and 2. Regarding exclusive control and logging, regardless of the structure of the data holding means for search, the data holding means for insertion 1, 2 and the data holding means 1, 2 for deletion, which constitute a file, they are grouped together as one object. Can handle.

Here, the lock will be described. In the lock method based on the two-phase rule, the write operation reserves the update (X) lock for the operation target, and the read operation reserves the reference (S) lock for the operation target, thereby reserving the target. From, operate. Table 1 shows the result when a certain transaction T1 has a certain kind of lock and another transaction T0 has already a certain kind of lock on the target.

[0057]

[Table 1]

In Table 1, ○ indicates that locks can coexist, and transaction T1 can lock the operation target. X indicates that the lock conflicts, and the transaction T1 suspends execution and waits until the lock causing the conflict is abandoned and the operation target can be locked.

In a transaction, if one lock is abandoned, the lock is not secured thereafter until the transaction ends. At the end of a transaction, any locks held by that transaction that have not been relinquished will be relinquished. That is, the transaction is divided into two phases: a process of securing a plurality of locks and a process of releasing the secured lock. This guarantees the serializability of the transaction.

When different types of operation objects are recognized during a transaction, for example, a database,
To permit operations on tables that make up the database, tuples that make up the tables, and so on, a multi-grain lock is required.
In other words, in a write operation on a certain object, IX or SIX indicating that there is an update intention is secured for a higher-order object, and then an X lock is secured for that object.
In a read operation on a certain object, an IS lock indicating that the user intends to refer is secured for a higher-order object, and then an S lock is secured for that object. When these intentional locks IX, SIX, and IS are added to Table 1, Table 2 is obtained. As in Table 1, ○ indicates that locks can coexist and x indicates that locks collide.

[0061]

[Table 2]

Hereinafter, in the present invention, READ COMM
A method for realizing the ITTED isolation level will be described. First, there are the following preconditions regarding locking. The lock on the table has one of the modes IS, IX, S, SIX, and X. Locks on tuples have either S or X mode.・ The S lock on the tuple can be released at any time after reference. -The X lock on the tuple cannot be released until the end of the transaction.・ Before and after accessing the file, the file is exclusively latched. The procedure is described according to the type of database operation.

(Search Operation) Lock the target table before starting the operation. 1. Latch the file. 2. Search for tuples that match the search conditions. 3. Apply S lock to the obtained tuple. 4. If the S lock cannot be applied, the latch to the file is released and the process returns to 1. 5. Add tuples to search results. 6. Unlock the tuple. 7. Unlatch to file.

(Insert Operation) IX lock the target table before starting the operation. 1. Latch the file. 2. Obtain a new tuple ID. 3. X lock the tuple. 4. Record tuple insertion in log. 5. Insert a tuple into the file. 6. Unlatch to file.

(Update Operation) IX lock the target table before starting the operation. 1. Latch the file. 2. Search for tuples that match the conditions to be updated. 3. X lock the tuples obtained. 4. If the X lock cannot be applied, the latch to the file is released and the process returns to 1. 5. Record tuple updates in the log. 6. Perform a tuple update. 7. Unlatch to file.

(Delete Operation) IX lock the target table before starting the operation. 1. Latch the file. 2. Search for tuples that match the conditions to be deleted. 3. X lock the tuples obtained. 4. If the X lock cannot be applied, the latch to the file is released and the process returns to 1. 5. Record tuple deletion in log. 6. Perform tuple deletion. 7. Unlatch to file. At the end of the transaction, the following processing is performed.

(Commit) 0. An IX lock has been placed on the target table before the change operation. 1. Record commit in log. 2. Release all locks requested by the committed transaction.

(Rollback (abort)) An IX lock has been placed on the target table before the change operation. 1. Record rollback in log. 2. Undo the data operation while referring to the log.
At that time, the necessary file is latched. 3. Release all locks requested by the rolled back transaction.

The embodiments of the present invention have been described centering on the database management system of the present invention. However, the present invention can also be implemented as a database management system as described as the processing procedure in the system. Further, the present invention can be implemented as a program for implementing these database management methods or functioning as each means of the database management system, or as a computer-readable recording medium on which the program is recorded.

An embodiment of a recording medium storing programs and data for realizing the function of database management according to the present invention will be described. As a recording medium, specifically,
A CD-ROM, a magneto-optical disk, a DVD-ROM, a floppy (registered trademark) disk, a flash memory, and various other ROMs and RAMs can be assumed. The functions of the system of each embodiment of the present invention described above are recorded on these recording media. By causing a computer to execute and record and distribute a program for realizing a database management function, the function is easily realized. Then, the program is read by the information processing apparatus by attaching the recording medium as described above to an information processing apparatus such as a computer, or the program is stored in a storage medium provided in the information processing apparatus, and if necessary, By reading, the database management function according to the present invention can be executed.

[0071]

According to the present invention, it is possible to maintain the ability to process high-level search requests provided by the data storage means for search at a high speed, and prevent the update performance from being lowered.

According to the present invention, the consistency between the data holding means for search, the data holding means for insertion 1, 2 and the data holding means 1, 2 for deletion in response to a database operation request asynchronously requested by a plurality of users. While responding to searches and updates quickly.

According to the present invention, data can be inserted at a high speed even while the data of the data holding means for insertion 1 or 2 is being reflected on the data holding means for search, and the response to the update request can be improved. it can.

According to the present invention, data can be deleted at a high speed even while the data in the data holding means for deletion 1 or 2 is being reflected on the data holding means for search, and the response to an update request can be improved. it can.

According to the present invention, for an existing database management system in which the retrieval is fast but the update is relatively slow, the update request can be made without any special change in the database management system. Response can be improved. Conversely, if the update response of the existing database management system is fast, the response to the search request can be improved while utilizing the update response performance by combining it with a high-speed data holding means.

According to the present invention, when there is a large amount of memory, remarkable improvement in update performance can be realized by using the memory as the data holding means for change.

According to the present invention, by using each data holding means corresponding to the characteristic of the attribute value, for example, for an attribute value having a small data amount, the memory is inserted into the data holding means 1, 2 and the data holding means for deletion. By using them as 1 and 2, the performance can be improved.

According to the present invention, when a time period during which the operation rate of the system is low is determined, the data throughput to the data holding means for retrieval is concentrated in that time period to increase the throughput of the entire system. be able to.

According to the present invention, by reducing the frequency at which update requests are issued to the data storage means for search, the frequency at which exclusive control is required can be reduced, and the throughput of the entire system can be increased.

According to the present invention, if the search data holding unit can execute a change operation on a plurality of tuples in a batch process to perform the operation at a high speed, the data transfer unit can execute the data at a high speed. Throughput can be increased.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating the functions of a database management system according to the present invention.

FIG. 2 is a diagram showing an example of a system configuration embodying the present invention.

FIG. 3 is a flowchart illustrating an example of a processing procedure of a database operation request processing unit.

FIG. 4 is a flowchart illustrating an example of a processing procedure of a database operation request processing unit.

FIG. 5 is a flowchart illustrating an example of a processing procedure of a database operation request processing unit.

FIG. 6 is a flowchart illustrating an example of a processing procedure of a data transfer unit.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Database operation request input means, 2 ... Transaction processing means, 3 ... Database operation request processing means, 4
... data holding means for search, 5 data holding means for insertion 1, 6 ... data holding means for insertion 2, 7 ... data holding means for deletion 1, 8 ... data holding means for deletion 2, 9, ...
Data transfer means, 10: server host, 11: CPU,
12: memory, 12a: program area, 12b: data area, 13: hard disk, 14: bus, 21: input terminal, 22: LAN.

 ──────────────────────────────────────────────────続 き Continued on the front page (72) Inventor Tetsuya Ikeda 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd. (72) Inventor Kazushige Asada 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd. (72) Inventor Hiroshi Takekawa 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd. (72) Yasushi Ogawa 1-3-6 Nakamagome, Ota-ku, Tokyo Ricoh Co., Ltd. F term (reference) 5B075 KK54 ND23 5B082 FA17 GA03 GB04 GB07

Claims (21)

[Claims] In a database management apparatus for managing a database, a data holding means for search which can process a data search operation at a high speed but can process a data change operation at a low speed, and has a high speed at a data change operation. A first and a second insertion data holding unit; a first and a second deletion data holding unit which process data change operations at a high speed; the first and the second insertion data holding unit; In order to reflect the data processed by the first and second data-for-deletion means, the search is performed from the first and second data-for-insertion means and the first and second data-for-deletion means. Transferring means for transferring data to a data holding means for the first and second data holding means for the first and second insertions according to the contents of the operation request. And a database operation request processing unit that executes the operation by distributing the data to the first and second data holding units for deletion and the data holding unit for search, respectively, between the data transfer unit and the database operation request processing unit. A database management device, comprising: transaction processing means for ensuring data consistency. 2. A database management apparatus for managing a database, comprising: a database operation request input unit for inputting a database operation request from a plurality of users; and a data search operation process is performed at a high speed, but a data change operation process is performed. A low-speed search data holding unit, a first and a second insertion data holding unit that processes data change operations at a high speed, and a high-speed first and second deletion data store that processes data change operations. Means for reflecting the data processed by the first and second data storage means for insertion and the first and second data storage means for deletion, Data transfer means for transferring data from the holding means and the first and second data holding means for deletion to the data holding means for search; An operation request to the database input to the source operation request input means, the first and second insertion data holding means and the first and second deletion data holding means in accordance with the contents of the operation request And a database operation request processing unit that distributes and executes each of the search data holding units, and a transaction processing unit that guarantees data consistency between the data transfer unit and the database operation request processing unit. A database management device, characterized in that: 3. The database management device according to claim 1, wherein, when data is requested to be inserted during insertion of the data of the first data storage unit for insertion into the data storage unit for search, A database management device, wherein the data requested to be inserted is inserted into a second data storage unit for insertion. 4. The database management apparatus according to claim 1, wherein when a data deletion request is received while data of the first data storage unit for deletion is being deleted from the data storage unit for search. A database management device, wherein the data requested to be deleted is inserted into a second data storage unit for deletion. 5. The database management device according to claim 1, wherein said search data holding unit, said first insertion data holding unit, said second insertion data holding unit, and Database management characterized in that one or more of the first data holding means for deletion and the second data holding means for deletion can constitute a database management system so that they can operate independently. apparatus. 6. The database management device according to claim 1, wherein said first insertion data holding unit, said second insertion data holding unit, and said first deletion data holding unit. And at least one of the second data-for-deletion holding means is in a memory. 7. The database management device according to claim 1, wherein the search data holding unit, the first insertion data holding unit, the second insertion data holding unit, The database management apparatus according to claim 1, wherein the first data holding means for deletion and the second data holding means for deletion have their forms selected according to the characteristics of attribute values. 8. The database management apparatus according to claim 1, wherein the data transfer unit holds the designated transfer time information, so that the data transfer unit can store the search data holding unit. The data management apparatus according to claim 1, wherein the data change operation can be delayed until a time specified in the transfer specification time information. 9. The database management apparatus according to claim 1, wherein the data transfer unit has a predetermined number of change operation requests to the search data holding unit to be executed by the data transfer unit. A database management device for executing a change operation every time. 10. The database management apparatus according to claim 1, wherein a change operation on a plurality of processing unit data in said search data holding unit is executed as a batch process. apparatus. 11. A data holding means for search which is fast in processing data search operation but is slow in processing data change operation, and first and second insertion data holding means which is fast in processing data change operation. And a database management method for processing and managing the database using high-speed first and second data storage means for deletion, wherein the data change operation is performed by a plurality of users. And the first and second data storage means for insertion and the first and second data to reflect the processed data.
Transferring data from the data holding means for deletion to the data holding means for search, and transmitting the input operation request to the database in accordance with the content of the operation request, the first and second data holding means for insertion And processing is separately performed between the first and second data holding means for deletion and the data holding means for search, and data consistency is guaranteed between the transfer of the data and the processing of the database operation request. A database management method, characterized in that: 12. The database management method according to claim 11, wherein when the data of the first data storage unit for insertion is inserted into the data storage unit for search, a request for data insertion is made. And inserting the data requested to be inserted into the data holding means for insertion. 13. The database management method according to claim 11, wherein when a request to delete data is issued while data in said first data storage for deletion is being deleted from said data storage for search, said second data is deleted. And inserting the data requested to be deleted into the data holding means for deletion. 14. The database management method according to claim 11, wherein the search data holding unit, the first insertion data holding unit, the second insertion data holding unit, Database management characterized in that one or more of the first data holding means for deletion and the second data holding means for deletion can constitute a database management system so that they can operate independently. Method. 15. The database management method according to claim 11, wherein the first insertion data holding unit, the second insertion data holding unit,
A database management method, wherein one or more of the first data holding means for deletion and the second data holding means for deletion are provided in a memory. 16. The database management method according to claim 11, wherein said search data holding unit, said first insertion data holding unit, said second insertion data holding unit, and The database management method according to claim 1, wherein the data holding means for deletion and the second data holding means for deletion have their forms selected according to the characteristics of attribute values. 17. The database management method according to claim 11, wherein designated time information for designating a time of the data transfer is held, so that the search data at the time of the data transfer is held. A database management method, wherein an operation of changing data to a holding unit is delayed until a time specified in the specified time information. 18. The database management method according to claim 11, wherein the data transfer comprises:
A database management method, wherein a change operation is executed every time a change operation request to the search data holding unit to be transferred reaches a predetermined number. 19. The database management method according to claim 11, wherein a change operation for a plurality of processing unit data in the data storage unit for search is executed as a batch process. . 20. A program for functioning as each means of the database management device according to any one of claims 1 to 10, or for executing the database management method according to any one of claims 11 to 19 . 21. A computer-readable recording medium on which the program according to claim 20 is recorded.