JP2643811B2 - Database reorganization method - 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 reorganization system, and more particularly, to a database reorganization system comprising a copy source database and a copy destination database comprising an index section and a data record section having a hierarchical structure. The present invention relates to a database reorganization method in a database system that reorganizes a database by copying.

[0002]

2. Description of the Related Art Conventionally, in this type of database reorganization method, a copy source database and a copy destination database are placed in an exclusive state (offline state) inaccessible from a business processing program, and data records are transferred from the copy source database to the copy destination database. The database was reorganized by copying.

[0003]

In the conventional database reorganization method described above, the copy source database and the copy destination database are set to an exclusive state in which the business processing program cannot access them, and data records are copied from the copy source database to the copy destination database. Therefore, when the database system is operated continuously for 24 hours, the copy source database cannot be put into an exclusive state in which the business processing program cannot access the database system. There was a problem that it could not be performed.

Therefore, a copy source database and a copy destination
Database is accessible from business processing programs
From the source database to the destination database in a shared state
When copying data records to
Data records between RAM and database reorganization scheme
As a shared control technology that prevents inconsistencies in
A database disclosed in Japanese Patent Application Laid-Open No. H4-175841.
A reorganization processing method has already been proposed. But this
In the conventional database reorganization processing method of
Data records are written to the database in order from the top.
The copy destination database has been reorganized.
There is not enough free space secured and copy after reorganization
Data records written to the destination database are logical
Can be written near the data record associated with
Data access performance is likely to deteriorate rapidly
There was a problem.

In view of the above, an object of the present invention is to provide a data
The database can be reorganized while the base is operating.
Database reorganization method
Prior to the source reorganization process,
Read data record and based on free space acquisition information
Index key placement after reorganization in the copy destination database
Simulate and index into the destination database according to the simulation results
Performs reorganization pre-processing that generates only the
Database to ensure that there is adequate free space in the database.
An object of the present invention is to provide a source reorganization method.

[0006]

According to the present invention, there is provided a database reorganization method comprising a copy source database and a copy destination database each comprising an index section and a data record section having a hierarchical structure. In a database system that reorganizes a database by copying data records to a destination, to set a free space in the copy destination database
User who acquires the free space acquisition information specified by the user in
Information obtaining means and the user information obtaining means.
User information storage means for storing the obtained free area acquisition information
Read a data record from the copy source database.
Outgoing data record information and the user information storage means
From the free space acquisition information stored in the
Simulate the layout of index keys after reorganization
Storage system for generating index information of the destination database
Database storage control information generating means for creating control information
Created by this database storage control information generation means
Storage control information storage means for storing the stored storage control information
And the storage control after initializing the copy destination database.
An index key based on the storage control information stored in the information storage means.
-A dummy record that does not have any information
After the code is temporarily stored in the copy destination database,
The copy destination data is deleted by deleting the me record.
Index information setting means that only generates the base index part
And Ranaru reorganization pre-processing means, the copy from the copy source data base and the copy destination database after creation of the index portion of the copy destination database by the reorganization pre-processing means wherein the work processing program accessible shared state From a database copy means for copying data records from a source database to the destination database
And a reorganization means to be.

[0007]

Next, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a database reorganization system according to one embodiment of the present invention. The database reorganization method of the present embodiment is based on
From the reorganization means 2, the user information storage means 4, the copy source database 6, the copy destination database 7, the storage control information storage means 8, the business processing program 11, and the free space acquisition information 12, Its main part is configured.

The reorganization pre-processing means 1 includes a user information obtaining means 3, a database storage control information generating means 5, and an index information setting means 9.

[0010] The reorganization means 2 includes a database copy means 10.

The copy source database 6 and the copy destination database 7 each include an index portion and a data record portion having a hierarchical structure. The data record includes a root record type stored and searched by an index key, and a logical relationship with the root record. And a record type arranged in the vicinity. Assuming that a set of one root record and data records logically related thereto is called a cluster, the copy source database 6 stores, for example, an index key value and a capacity (converted into the number of blocks) shown in FIG. ) Are stored as shown in FIG. In the state of FIG. 3, the number of blocks in one index control area indicating the number of blocks to which one piece of index information is set is three.
Therefore, the index key value and the starting block number of the index part are as shown in the figure.

FIG. 4 shows the free space acquisition information 12, and how many blocks an index key is set as information for simulating (simulating) the index information to be set (hereinafter, referred to as an index key).
5 is set as one index control area block number, and how many index keys are set to obtain an empty area (hereinafter, referred to as an index control area block number).
2 is input as the number of free area blocks indicating the number of free area blocks to be taken at one time.

FIG. 5 is a diagram exemplifying a storage control information record stored in the storage control information storage means 8 . Each storage control information record is composed of an identifier, an index key value, and a capacity (converted to the number of blocks). The storage control information of an identifier (= “D”) indicating that the information is area information for data storage. The index key value is stored in the record, but the index key value is not stored in the storage control information record of the identifier (= “S”) indicating free space information.

FIGS. 6A to 6C are diagrams illustrating an example of a procedure for setting the index information by the index information setting means 9. FIG.
(A) shows the setting of the first index key value, (b) shows the setting of the second index key value, and (c) shows the setting of the third index key value.

FIG. 7 is a diagram exemplifying a storage state and an index portion in the copy destination database after the reorganization.

Referring to FIG. 8, the processing of the database storage control information generating means 5 is a counter initialization step S
101, cluster input step S102, cluster end determination step S103, index key value accumulation step S104, block number addition step S105, one index control area block number determination step S106, storage control information record for data storage And output step S
107 and storage control information counter adding step S108
Block number counter clear step S109, index key value stack clear step S110, free area setting frequency determining step S111, free area storage control information record creation and output step S112, and storage control information counter clear step S113. Consists of

Referring to FIG. 9, the processing of the index information setting means 9 is a copy destination database initialization step S201.
And the next storage block pointer 1 initialization step S202
, A storage control information record input step S203, a storage control information record end determination step S204, an identifier determination step S205, a dummy record creation, storage and deletion step S206, and a next storage block pointer addition step S207.

Next, the operation of the database reorganization system according to the present embodiment configured as described above will be described.

When the database needs to be reorganized, the copy source database 6 is placed in an exclusive state inaccessible from the business processing program 11 while the copy source database 6 remains in a shared state accessible from the business processing program 11. The reorganization pre-processing means 1 is activated.

Then, the user information acquisition means 3 acquires and analyzes the free space acquisition information 12 and stores it in the user information storage means 4.

Next, the database storage control information generating means 5 initializes a block number counter (not shown), a storage control information counter (not shown), and an index key value stack (not shown) (step S101). Then, clusters are input from the copy source database 6 in the order of index key values (step S102).

When a cluster is input (step S10)
3), the database storage control information generating means 5
The index key value of the root record of the cluster is accumulated in the index key value stack (step S104), the record size of the cluster is integrated, and the converted block number value is added to the block number counter (step S105).

Subsequently, the database storage control information generating means 5 determines whether or not the block number counter is equal to or greater than the number of one index control area blocks (step S106).
If the block number counter is less than one index control area block number, control is returned to step S102.

If the number-of-blocks counter is equal to or greater than one index control area block number (YES in step S106), the database storage control information generating means 5 sets the first (earliest) value stored in the index key value stack. The index key value and the number of blocks of the block number counter are output to the storage control information storage means 5 as a storage control information record together with an identifier (= "D") indicating that the information is area information for data storage (step S107). Thereafter, the database storage control information generating means 5 adds one to the storage control information counter (step S108), clears the block number counter to 0 (step S109), and clears the index key value stack (step S110). .

Next, the database storage control information generating means 5 compares the storage control information counter with the free area setting frequency (step S111). If the storage control information counter is less than the free area setting frequency, the processing proceeds to step S102. Return control.

If the storage control information counter is equal to or higher than the free area setting frequency (Yes in step S111), the database storage control information generating means 5 determines the number of free area blocks as an identifier (= “S ) Is output to the storage control information storage means 8 as a storage control information record (step S112). Thereafter, the database storage control information generating means 5 clears the storage control information counter to 0 (step S113), and proceeds to step S10.
Return control to 2.

The database storage control information generating means 5 comprises:
Steps S102 to S113 are repeated, and step S1
When the cluster is no longer input from the copy source database 6 in 03, the processing is terminated.

After the processing of the database storage control information generating means 5 ends, the index information setting means 9 initializes the copy destination database 7 (step S201), and initializes the next storage block pointer to 1 (step S20).
2) The storage control information record is input from the storage control information storage means 8 (step S203).

If a storage control information record is input (No in step S204), the index information setting means 9 determines an identifier (step S205), and an identifier indicating that the identifier is data storage area information (step S205). = “D”), the index key value in the storage control information record is itself
It sets the index information in the index section by storing in set dummy records that do not have the information to the copy destination database 7, only create index portion by removing the dummy records immediately (step S206). This
Here, the deletion of the dummy record is based on the copy destination database.
To create an index part once stored in source 7
Dummy records that have completed the target occupy the physical area
This is to eliminate the adverse effects. If it is determined in step 205 that the identifier is free space information (= “S”), the index information setting unit 9 skips step S206.

Next, the index information setting means 9 adds the number of blocks in the storage control information record to the next storage block pointer (step S207), and returns the control to step S203.

The index information setting means 9 determines in step S203
Steps S207 to S207 are repeated to create only the index part in the copy destination database 7. When the storage control information record in the storage control information storage unit 8 is exhausted in step S204, the process is terminated.

After the processing of the reorganization preprocessing means 1 is completed, the copy source database 6 remains in the shared state accessible from the business processing program 11 and the copy destination database 7 remains in the shared state accessible from the business processing program 11. And the reorganization means 2 is activated.

Then, the database copy means 8 inputs the data record from the copy source database 6 and stores it in the copy destination database 7. At this time, the data records to be stored are stored according to the index section already set by the reorganization pre-processing means 1, for example, in the storage state shown in FIG. For the data records stored in the copy destination database 7, the copy source database 6
To turn on a copied flag (not shown) in the data record to uniquely determine the data record to be accessed. Incidentally, as disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 4-1755841, by adding a display of reorganization to the index portion, it is possible to prevent inconsistency in the data record.

During the processing of the database copying means 8, the business processing program 11 accesses the copy destination database 7 for copied data records and the copy source database 6 for uncopied data records.

[0035]

As described above, the present invention provides a database
Database reorganization processing in the operating state of the database
Database reorganization method
Prior to the data reorganization process,
Data record is read and based on the free space acquisition information.
Mimics the placement of index keys after reorganization in the destination database
Simulate the index part in the copy destination database according to the simulated result.
Performs reorganization pre-processing to generate
A reasonable free space for
Data inserted into the destination database after reorganization
Because records are stored in indexed blocks,
Data access performance can be slowed down
Has the effect.

[0036]

[0037]

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a database reorganization method according to an embodiment of the present invention.

FIG. 2 is a diagram exemplifying an index key value and a capacity of a cluster in a copy source database in FIG. 1;

FIG. 3 is a diagram exemplifying a storage state and an index unit in a copy source database in FIG. 1;

FIG. 4 is a diagram showing an example of free area acquisition information in FIG.

FIG. 5 is a diagram showing an example of the contents of a storage control information storage means in FIG. 1;

6A and 6B are diagrams exemplifying a procedure for setting index information by the index information setting means in FIG. 1, wherein FIG. 6A is a diagram for setting a leading index key value, FIG. 6B is a diagram for setting a second index key value, (C) shows the time when the third index key value is set.

FIG. 7 is a diagram exemplifying a storage state and an index part of the copy destination database in FIG. 1 after reorganization;

FIG. 8 is a flowchart showing a process of a database storage control information generating means in FIG. 1;

FIG. 9 is a flowchart showing the processing of the index information setting means in FIG. 1;

[Explanation of symbols]

 1 Reorganization preprocessing means 2 Reorganization means 3 User information acquisition means 4 User information storage means 5 Database storage control information generation means 6 Copy source database 7 Copy destination database 8 Storage control information storage means 9 Index information setting means 10 Database Copying means 11 Business processing program 12 Free space acquisition information

Claims (4)

(57) [Claims] A copy source database and a copy destination database each comprising an index section and a data record section having a hierarchical structure, and a database is reorganized by copying data records from the copy source database to the copy destination database. In the database system to be used, it is useful to set the free area of the copy destination database.
User information for acquiring the free space acquisition information specified by the user
Acquiring means and the user information acquired by the user information acquiring means.
User information storage means for storing free space acquisition information;
Reads data records from the source database
Data record information and stored in the user information storage means
From the obtained free space acquisition information and the copy destination database
Simulate the index key arrangement after reorganization in
Storage control information for generating database index information
Database storage control information generating means for creating
The case created by the database storage control information generation means
Storage control information storage means for storing delivery control information;
After initializing the destination database, the storage control information storage
Set the index key value based on the storage control information stored in the column
A dummy record that does not have any information
Once stored in the destination database,
By deleting the database, the destination database can be searched.
And reorganization pre-processing means comprising an index information setting means for performing only the generation of引部, the copy source database and the copy destination database after creation of the index portion of the copy destination database by the reorganization pre-processing means wherein Day of copying data records accessible shared state from the business processing program from the copy source database to the copy destination database
A database reorganization method comprising a database reorganization unit comprising a database copy unit . 2. A cluster in which a root record stored and retrieved in the copy source database and the copy destination database by an index key and a record which is logically related to the root record and arranged near the root record form a cluster. Item 2. The database reorganization method according to Item 1. 3. The free area acquisition information indicates the number of one index control area block indicating how many blocks an index key is set and how many index keys are set to obtain a free area. A claim comprising: a free area acquisition frequency; and a free area block number indicating the number of free area blocks taken at one time.
1. The database reorganization method according to 1. Wherein said storing control information, an identifier indicating the area information or free space information for data storage, an index key value and the database reorganization method according to claim 1, wherein comprising a capacitor.