AU2002256151A1 - System and method for reorganizing stored data - Google Patents

Description

SYSTEM AND METHOD FOR REORGANIZING STORED DATA

Technical Field

The described system and method are generally related to information processing

environments and systems and methods for database or file accessing. More specifically, the

described system and method are related to systems and methods for improving the

reorganization of a tablespace or index.

Background

Computers are powerful tools for storing, managing and providing access to vast

amounts of information. Computer databases are one common mechanism for storing

information on a computer while providing access to users. Common computer

implementations of databases store data and indexes in various files or object.

Typically, users do not have direct access to the objects in which the data and/or

indexes comprising a database are stored. Users are often provided indirect access to the data

and indexes via a database management system ("DBMS"), or an application communicating

with a DBMS. A DBMS is responsible for responding to requests from users or applications

to change, update, delete and insert data into the physical objects. In this way, the DBMS

acts as a buffer between the end-user and the physical data storage mechanism, thereby

shielding the end user from having to know or consider the underlying hardware-level details

of the table he is using.

There are several common database management systems including, for example,

DB2 which employs tablespace and index objects to store and access data. Another example

of a common DBMS implementation is IMS which employs database and index objects to store and access data.

In a typical database environment, rows of user data resides in tables which are

maintained in data objects such as databases or tablespaces. Each object storing user data

may have one or more indexes. Each index facilitates access to rows of the table according to

a key. The key of an index is typically data from one or more columns of the table. The rows

of data are available to batch and online applications for reading, updating, deleting and

inserting new data. When a row of data is inserted or deleted, a corresponding insertion or

deletion is performed on all associated indexes. When a key column is updated, all

corresponding indexes are also updated.

Typical tables and indexes may include thousands of records. In many DBMS's, all

changes, updates, deletions and insertions to the objects are recorded to a log file. The log

function is one of the busiest functions in a DBMS due to the large number of records and the

high volume of changes being made to objects. A typical DBMS log function also allows for

a log exit. Namely, before the DBMS writes each log record, it calls a log exit routine and

passes the address of the log record to the routine.

Over time, changes, additions and deletions from a table and/or index may result in an

inefficient organization of the stored data, and may affect the ability of the DBMS to timely

respond to requests from end-users and applications. To maintain efficient data storage and

access, utilities have been developed to reorganize data and index objects. Such utilities may

be periodically executed to correct the inefficient organization of data caused by the

processing of requests since the last time the reorganization utility was executed.

Reorganization utilities are employed periodically because of the time and resources required

to perform the reorganization of data.

While a reorganization utility is executing, batch and online applications which require access to the data and/or index objects being reorganized may be executing

concurrently. For this reason, reorganization utilities typically examine and reorganize the

subject objects in two phases. In the first phase, the subject object is reorganized to account

for all changes which have occurred up to the execution of the reorganization utility.

In the second phase, the typical reorganization utility accounts for all changes which

have occurred during the execution of the reorganization utility. This is accomplished by

reviewing all log file records reflecting changes requested by the concurrent batch and online

applications. Before completing the reorganization, the utility processes all of the changes

written to the log file, thereby providing an up to date reorganization of the subject data or

index object.

A typical DBMS environment is illustrated in Figure 1. As shown, the environment

includes a database 110 for maintaining and allowing access to stored information. Database

110 includes at least one data object 112 for storing rows and columns of data. Database 110

preferably includes one or more indexes 114 and 116 associated with data object 112 to assist

in accessing the data stored therein. Of course, indexes 114 and 116 are optional, and data

object 112 is not required to have any index.

Access to database 110 is provided by Database Management System ("DBMS") 120.

DBMS 120 enables user 130 to access database 110. DBMS 120 also enables user 150 to

access database 110 indirectly through application 140. DBMS 120 includes routines for

reading, adding, deleting and changing the data stored in database 110. DBMS 120 also

includes at least one routine for logging all changes made to any object managed by DBMS

120. The logging function may utilize a log database 122 embodied as a data object 124 and

an index object 126. In addition to routines for logging changes, DBMS 120 further includes

utilities for maintaining the integrity of the data stored in data object 112 and indexes 114 and 116. Certain utilities may be used to rebuild the files or objects within database 110 in the

event they become corrupted. Other utilities, specifically a reorganization utility may be used

to rearrange the data stored in database 110 for more efficient access. The reorganization

utility may operate on data object 112, index 114, index 116, and any combination thereof.

Referring now to Figure 2, there is depicted the steps that a conventional DBMS log

routine executes each time a data or index entry is added, deleted or modified. At step 210 a

log record is created in a log file. The log file contains changes made by the DBMS to data

and/or index objects. The log record identifies the affected data or index object, identifies the

record of the affected file and describes the type of activity that resulted in a change to the

record. At step 212, a log exit routine is called. The log exit routine is called prior to the

writing of the log record, and the address of the log record is passed as part of the call. At step

214, the log record is actually written to the log file.

Referring generally to Figures 3A and 3B, there is depicted a block diagram

illustrating the steps that a conventional reorganization utility performs to more efficiently

store data. The steps are collectively referred to by reference numeral 300. Although

conventional reorganization utilities may operate on both data and index objects, Figures 2,

3A, and 3B are described in terms of reorganizing a data object. Of course, analogous steps

are performed when reorganizing an index.

Reorganization utility 300 operates in two phase. During the first phase, depicted in

Figure 3A, the utility individually copies each record from the data object as it exists at the

beginning of the reorganization. During the second phase, depicted in Figure 3B, the

reorganization utility accounts for any changes that are made to the data object while

processing the first phase. Such changes may be requested by users, online applications or

batch applications that require access to the data object concurrently with processing the first phase of the reorganization.

Referring now to Figure 3A, the steps of the first phase of a conventional

reorganization utility are depicted. At step 310, the reorganization utility creates an empty

"shadow" data object based on the format of the real data object to be reorganized. Each

record of the real data object is read at step 312. As illustrated by decision block 314, if

attempting to read a record from the real data object at step 312 results in an End-of-File

condition, the reorganization utility begins the second phase of processing. If a record is

successfully read at step 312, the record is written to the shadow data object at step 316.

Referring now to Figure 3B, the steps of the second phase of a conventional

reorganization utility are depicted. Upon entering the second phase of processing, at step

318, the reorganization utility searches for the first record in the log file that pertains to a

record of the data object being reorganized, where the logged change occurred after the

reorganization utility was invoked. In subsequent iterations, step 318 will search for the next

record in the log file that pertains to a record of the data object being reorganized. At step

320, the record is read from the log file.

As illustrated by decision block 322, if attempting to read a record from the log file at

step 320 results in an End-of-File condition, the reorganization utility completes the

reorganization process by performing step 326. If a log file record is successfully read at step

320, the change described by the log file record is applied to the shadow index at step 324,

and processing is directed back to step 318. After all concurrent changes to the data object

have been applied, the newly reorganized shadow data object is renamed to become the real

data object at step 326, thereby allowing access to the reorganized data object.

Consequently, a need exists for an improved method and system for reorganizing data

that enables a reorganization utility to operate more efficiently than conventional reorganization utilities. Specifically, a need exists for a method and system that reduces the

processing related to effecting changes that are made to a file while it is being reorganized.

Summary

A method for reorganizing data is described. The method includes the steps of reading

each record of a source file and writing each record to a destination file. The method also

includes the step of creating a log file containing selected log records. Each log record is

associated with a change to be made to the destination file. The method further includes the

steps of reading each log record of the log file and processing each record of the log file to

effect the associated change to the destination file. The method still further includes the step

of replacing the source file with the destination file.

A method for logging changes by a database management system is also described.

The method includes the steps of identifying a change to be logged and creating a log record

based on the change. The method also includes the step of determining whether the change

affects a reorganization process. If the change affects the reorganization process, the log

record is stored in a first log file. The first log file records only selected changes. The

method further includes the step of storing the log record in a second log file. The second log

file records all changes.

An first apparatus for reorganizing data is also described. The apparatus includes a

means for reading each record of a source file and a means for writing each record to a

destination file. The apparatus also includes a means for creating a log file containing

selected log records. Each log record is associated with a change to be made to the

destination file. The apparatus further includes a means for reading each log record of the log

file and a means for processing each record of the log file to effect the associated change to the destination file. The apparatus still further includes a means for replacing the source file

with the destination file.

A second apparatus is described for reorganizing data. The second apparatus includes

a processor, and a memory connected to the processor. The memory stores a program to

control the operation of processor to carry out the steps of the described method for

reorganizing data.

An article of manufacture is also described. The article of manufacture is a computer-

readable storage medium encoded with processing instructions for implementing the

described method for reorganizing data.

The objects, features and advantages of the disclosed method and system are readily

apparent from the following description of the preferred embodiments when taken in

connection with the accompanying drawings.

Brief Description of the Drawings

For a more complete understanding of the disclosed method and system and the

advantages thereof, reference is now made to the following description taken in conjunction

with the accompanying drawings in which like reference numbers indicate like features and

wherein:

Figure 1 is a schematic block diagram illustrating a typical DBMS environment;

Figure 2 is a functional flow diagram illustrating the primary steps employed by a

conventional DBMS log routine;

Figures 3A-3B are a functional flow diagram illustrating the primary steps typically

employed by a conventional reorganization utility;

Figure 4 is a schematic block diagram illustrating the environment of an embodiment of the described method and system;

Figures 5A-5B are a functional flow diagram illustrating the primary steps of an

embodiment of a reorganization utility according to the described reorganization method; and

Figure 6 is a functional flow diagram illustrating the primary steps of an embodiment

of a DBMS log routine according to the described reorganization method.

Detailed Description

Environment

The disclosed method of reorganizing data preferably operates in an environment such

as that illustrated in Figure 4. As shown, the environment includes a database 410

comprising at least one data object 412 and one or more indexes 414 and 416 associated with

data object 412 to assist in accessing the data stored therein. Of course, indexes 414 and 416

are optional, and data object 412 is not required to have any index.

Access to database 410 is provided by Database Management System ("DBMS") 420.

DBMS 420 enables user 430 to access database 410. DBMS 420 also enables user 450 to

access database 410 indirectly through application 440. DBMS 420 includes routines for

reading, adding, deleting and changing the data stored in database 410.

A logging routine of DBMS 420 logs all changes made to any object managed by

DBMS 420 in a log database 422 embodied as a data object 424 and an index object 426. In

addition to routines for logging changes, DBMS 420 further includes utilities for maintaining

the integrity of the data stored in data object 412 and indexes 414 and 416. Certain utilities

may be used to rebuild the files or objects within database 410 in the event they become

corrupted. Other utilities, specifically a reorganization utility may be used to rearrange the

data stored in database 410 for more efficient access. The reorganization utility may operate

on data object 412, index 414, index 416, and any combination thereof. The reorganization

utility utilizes log database 427 including log data object 428 and log index object 429. Log

database 427 stores a subset of changes logged in log database 422.

Generally referring now to Figures 5A and 5B, there is depicted a block diagram

illustrating the steps of one embodiment of a reorganization utility according to the present

application. The steps are collectively referred to by reference numeral 500. Although the disclosed reorganization utility may operate on both data and index objects, Figures 5 A, 5B

and 6 are described in terms of reorganizing a data object. Of course, analogous steps are

performed when reorganizing an index and are considered to be within the scope of the

described method and system.

Improved reorganization utility 500 operates in two phases. During the first phase,

depicted in Figure 5A, the improved reorganization utility establishes a program call to be

used by the DBSM log routine and individually copies each record from the data object as it

existed at the beginning of the reorganization. During the second phase, depicted in Figure

5B, the improved reorganization utility utilizes an duplicated subset of log records to account

for changes made to the data object during the first phase.

Referring now to Figure 5A, the steps of the first phase of the improved

reorganization utility are depicted. At step 510, the improved reorganization utility creates an

empty shadow data object based on the format of the real data object to be reorganized. At

step 512, the improved reorganization utility establishes a program call to be used by the

DBMS log routine, described in more detail with reference to Figure 6. The established

program call examines the log records and makes a copy of log records associated with

changes to the object being reorganized. These selected log records may be stored to a log

file, but are preferably stored in memory to improve the efficiency of the reorganization

utility. In the event a threshold is reached in memory utilization, the log records may be

stored in DASD. Accordingly, the memory and duplicate log file will only include the log

records that are to be processed by the reorganization utility, thereby more efficiently

processing the second phase.

Each record of the real data object is read at step 514. As illustrated by decision block

516, if attempting to read a record from the real data object at step 514 results in an End-of- File condition, the improved reorganization utility begins the second phase of processing. If a

record is successfully read at step 514, the record is written to the shadow data object at step

518.

Referring now to Figure 5B, the steps of the second phase of the improved

reorganization utility are depicted. Upon entering the second phase of processing, at step

520, the reorganization utility accesses the area in memory and/or the duplicate log file

containing the records relevant for the second phase of processing, and reads the first log

record. In subsequent iterations, step 520 will read the next record in the log file.

At decision block 522, the improved reorganization utility determines whether step

520 resulted in an End-of-File condition. If so, the reorganization utility continues processing

at step 526. If a log file record is successfully read at step 520, the change described by the

log file record is applied to the shadow index at step 524, and processing is directed back to

step 520. After all logged changes to the data object have been applied, the program call is

removed at step 526 and the newly reorganized shadow data object is renamed to become the

real data object at step 528, thereby allowing access to the reorganized data object.

Referring now to Figure 6, there is depicted the steps that an improved DBMS log

routine executes each time a data or index entry is added, deleted or modified, according to

the described method and system. At step 610 a log record is created in a log file that

contains changes made by the DBMS to data and/or index objects. At step 612, a log exit

routine is called, and the address of the log record is passed as part of the call. As shown by

decision block 614, if the program call has been established by the improved reorganization

utility, as previously discussed with reference to step 512 of Figure 5 A, the log routine

processes step 616.

At step 616, if the log record represents a change to a file currently being reorganized, the log record is copied to a duplicate log file for use in the second phase of the

reorganization. If no program call has been established, or upon completing processing of

step 616, the log record is written to the conventional log file at step 618.

From the above description, those skilled in the art will perceive improvements,

changes and modifications in the disclosed method and system. Such improvements, changes

and modifications within the skill of the art are intended to be covered by the appended

claims.

Accordingly, it is to be understood that the drawings and description in this disclosure

are proffered to facilitate comprehension of the disclosed method and system, and should not

be construed to limit the scope thereof. It should be understood that various changes,

substitutions and alterations can be made without departing from the spirit and scope of the

disclosed method and system as defined solely by the appended claims.

Claims (27)

What is claimed is:

1. A method for reorganizing data, comprising:

reading each record of a source file;

writing each record to a destination file;

creating a log file containing selected log records, each log record associated with a

change to be made to the destination file;

reading each log record of the log file;

processing each record of the log file to effect the associated change to the destination

file; and

replacing the source file with the destination file.

2. A method according to claim 1 wherein the source file is an index file.

3. A method according to claim 1 wherein the source file is a data file.

4. A method according to claim 1 wherein the step of creating a log file is performed in

accordance with instructions of a DBMS log routine.

5. A method according to claim 4 wherein the log file contains a subset of all records

processed by the DBMS log routine.

6. A method according to claim 4 wherein the log file records are selected based on a

program call established by a reorganization utility.

7. A method according to claim 6 wherein the program call is removed prior to

termination of the reorganization utility.

8. A method for logging changes by a database management system, comprising:

identifying a change to be logged;

creating a log record based on the change;

determining whether the change affects a reorganization process;

storing the log record in a first log file recording selected changes if the change affects

the reorganization process; and

storing the log record in a second log file recording all changes.

9. A method according to claim 8 wherein the first log file resides in virtual storage.

10. A method according to claim 8 wherein the first log file resides in dataspace.

11. A method according to claim 8 wherein the first log file resides in hiperspace.

12. A method according to claim 8 wherein the first log file resides in DASD.

13. An apparatus for reorganizing data, comprising:

means for reading each record of a source file;

means for writing each record to a destination file;

means for creating a log file containing selected log records, each log record

associated with a change to be made to the destination file;

means for reading each log record of the log file;

means for processing each record of the log file to effect the associated change to the

destination file; and

means for replacing the source file with the destination file.

14. An apparatus according to claim 13 wherein the source file is an index file.

15. An apparatus according to claim 13 wherein the source file is a data file.

16. An apparatus according to claim 13 wherein the log file is created in accordance with

instructions of a DBMS log routine.

17. An apparatus according to claim 16 wherein the log file contains a subset of all

records processed by the DBMS log routine.

18. An apparatus according to claim 16 wherein the log file records are selected based on

a program call established by a reorganization utility.

19. An apparatus according to claim 18 wherein the program call is removed prior to

termination of the reorganization utility.

20. An apparatus for reorganizing data, comprising:

a processor;

a memory connected to said processor storing a program to control the operation of

said processor;

the processor operative with the program in the memory to:

read each record of a source file;

write each record to a destination file;

create a log file containing selected log records, each log record associated

with a change to be made to the destination file;

read each log record of the log file;

process each record of the log file to effect the associated change to the

destination file; and

replace the source file with the destination file.

21. An apparatus according to claim 20 wherein the source file is an index file.

22. An apparatus according to claim 20 wherein the source file is an data file.

23. An apparatus according to claim 20 wherein the processor is further operative with the

program in the memory to create the log file in accordance with instructions of a DBMS log

routine.

24. An apparatus according to claim 23 wherein the log file contains a subset of all records processed by the DBMS log routine.

25. An apparatus according to claim 20 wherein the processor is further operative with the

program in the memory to select the log file records based on a program call established by a

reorganization utility.

26. An apparatus according to claim 23 wherein the processor is further operative with the

program in the memory to remove the program call prior to termination of the reorganization

utility.

27. A computer-readable storage medium encoded with processing instructions for

implementing a method for reorganizing data, the processing instructions for directing a

computer to perform the steps of:

reading each record of a source file;

writing each record to a destination file;

creating a log file containing selected log records, each log record associated with a

change to be made to the destination file;

reading each log record of the log file;

processing each record of the log file to effect the associated change to the destination

file; and

replacing the source file with the destination file.