Data partitioning by co-locating referenced and referencing records

• US 6,092,061 A
• Filed: 08/15/1997
• Issued: 07/18/2000
• Est. Priority Date: 08/15/1997
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Choy, David Mun-Hien
• Primary Examiner(s): HOMERE, JEAN RAYMOND
• Application Number:
  US08/911,717
• Time in Patent Office: 1,068 Days
• Field of Search: 707/3, 707/10, 707/100, 707/102, 707/2, 707/101, 395/200.31, 395/200.38, 395/200.47, 395/200.45
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2282 Tablespace storage structures; Management thereof
• CPC Class Codes: G06F 16/278 Data partitioning, e.g. horizontal or vertical partitioning
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A partitioning scheme for co-locating records of a table with records they reference. In a distributed database system, records of one table or relation may be stored at different locations. In order to optimize database query processing the present invention partitions records of a table by reference. A record in a table may reference another record in the table or in another table. The present invention stores (or co-locates) the referencing record with the referenced record.

Claims

1. In a computer system having a plurality of storage locations for storing data records, a method of partitioning a group of data records in which at least one data record includes a reference to at least one other data record which in turn includes a reference to at least a third data record stored in one of the plurality of storage locations, comprising:

identifying the at least one data record of the group of data records referencing the at least one other data record, the at least one other data record in turn referencing and being co-located with the at least a third data record;

identifying the one of the plurality of storage locations where the at least one other data record is stored; and

storing the at least one data record in the identified one of the plurality of storage locations.

2. The method according to claim 1, wherein the group of data records comprises tuples of a database relation.

3. The method according to claim 2, the computer system having a plurality of database systems networked together, each database system being located at a storage location and including a database management system (DBMS).

4. The method according to claim 3, wherein the group of data records comprises tuples of a first table and the at least one other data record is a tuple of a second table.

5. The method according to claim 3, wherein the group of data records comprises tuples of a first table and the at least one other data record is a tuple of the first table.

6. The method according to claim 1, wherein identifying includes:

maintaining an index of references to referenced data records, in which each reference is associated with a storage location of a referenced data record; and

identifying the one of the plurality of storage locations where the at least one other data record referenced by the at least one data record is stored by locating the at least one other data record in the index of referenced data records.

7. The method of claim 1, wherein identifying comprises computing a function on the reference.

8. In a database system having a plurality of storage locations, a method of partitioning a relation having at least one tuple that includes a reference to at least one other tuple which in turn includes a reference to at least a third tuple stored in one of the plurality of storage locations, comprising:

identifying the at least one tuple referencing the at least one other tuple. the at least one other tuple referencing and being co-located with the at least a third tuple;

identifying the one of the plurality of storage locations where the at least one other tuple referenced by the at least one tuple is stored; and

storing the at least one tuple in the identified one of the plurality of storage locations.

9. The method according to claim 8, wherein the at least one other tuple is part of another relation.

10. The method according to claim 8, wherein the at least one other tuple is part of the relation.

11. The method according to claim 8, wherein identifying includes:

maintaining an index of references to referenced tuples, in which each reference is associated with a storage location of a referenced tuple; and

identifying the one of the plurality of storage locations where the at least one other tuple referenced by the at least one tuple is stored by locating the at least one other tuple in the index of referenced tuples.

12. The method of claim 8, wherein identifying includes computing a function on the reference.

13. A method of inserting a record into a group of records in one of a plurality of storage locations where a field of the record includes a reference to a record of a first group of records and a field of the record of the first group in turn includes a reference to a record of a second group of records and where the record of the second group of records may be stored in one of the plurality of storage locations, the method comprising:

identifying the record referencing the record of a first group of records, the record of the first group of records in turn referencing and being co-located with the referenced record of the second group of records;

identifying the one of the plurality of storage locations where the record of the second group of records is stored; and

inserting the record into the first group of records by storing the record in the identified one of the plurality of storage locations.

14. The method according to claim 13, wherein the first and second group of records are databases.

15. The method according to claim 13, wherein the first and second group of records are relations, the records are tuples, and the field is an attribute of the first relation.

16. The method according to claim 15, wherein the attribute of the first relation is a foreign key that references a primary key of the second relation.

17. The method according to claim 13, wherein identifying includes:

maintaining an index of references to referenced records of the second group of records, in which each reference is associated with a storage location of a referenced record; and

identifying the one of the plurality of storage locations where the record of the second group of records is stored by locating the record of the second group of records in the index.

18. The method according to claim 13, wherein identifying includes computing a function on the reference.

19. A method of inserting a plurality of records into a group of records in one of a plurality of storage locations where a field of at least one of the plurality of records includes a reference to a record of a first group of records and a field of the record of the first group of records includes a reference to a record of a second group of records and where the record of the second group of records may be stored in one of the plurality of storage locations, the method comprising:

identifying the at least one of the plurality of records referencing the record of the first group of records, the record of the first group of records referencing and being co-located with the referenced record of the second group of records;

identifying the one of the plurality of storage locations where the record of the second group of records is stored; and

inserting the at least one of the plurality of records into the first group of records by storing the at least one of the plurality of records in the identified one of the plurality of storage locations.

20. A method according to claim 19 further comprising inserting any of the plurality of records which do not include a field which references a record of a second group of records into the first group of records by storing into one of the plurality of storage locations as a function of an attribute of the record.

21. A method of inserting a record into a group of records in one of a plurality of storage locations where a field of the record includes a reference to a record in a first group of records and a field of the record in the first group of records in turn includes a reference to a record in a second group of records, and where the record in the second group of records may be stored in one of a plurality of storage locations, the method comprising:

identifying the record referencing the record in the first group of records, the record of the first group of records referencing and being co-located with the referenced record of the second group of records;

identifying the one of the plurality of storage locations where the record in the second group of records is stored; and

inserting the record into the group of records by storing the record in the identified one of the plurality of storage locations.

22. The method according to claim 21, wherein the group of records is comprised of at least one database relation.

23. The method according to claim 22, wherein the field of the record is an attribute of at least one database relation.

24. The method according to claim 23, wherein the attribute of the at least one database relation is a foreign key that references a primary key of at least one database relation in the group of records.

25. The method according to claim 21, wherein identifying includes:

maintaining an index of referenced records in the group of records in which each reference is associated with a storage location of the referenced record; and

identifying the one of the plurality of storage locations where the record in the group of records is stored by locating the record in the group of records in the index.

26. The method according to claim 21, wherein identifying includes computing a function on the reference.