Implementing Siebel Business Applications on DB2 UDB for z/OS > Understanding Siebel Partitioning >

About Partitioning Keys

When a table is created on a partitioned table space, the table is assigned a partitioning key that is composed of one or more columns. Value ranges are assigned to each partition based on value ranges within the partitioning key. The value ranges determine which partition a particular row is assigned to.

How key values are specified for each table partition depends on whether the table uses table-controlled or index-controlled partitioning. With index-controlled partitioning, a partitioning index specifies the partitioning key and the key value ranges that determine how data is partitioned. With table-controlled partitioning, the partition key and key value ranges are contained in the table definition, so a partitioning index is not required. The Siebel 8.0 partitioned tables use table-controlled partitioning by default.

An example of partitioning in shown in Figure 2. You can divide a large table such as S_ORG_EXT (which holds a list of new accounts) to store records of names beginning with letters A-L in one partition and records of names beginning with letters M-Z in another partition. In this example, the partitioning key is the Name column and a partitioning index defines the key ranges (A-L; M-Z) for the partitioning column. A key range must be specified for each partition.

Figure 2. Partitioning S_ORG_EXT Table on DB2 for z/OS

NOTE:  For more information about partitioning table spaces on DB2, consult your IBM documentation.