Upgrade Guide for DB2 UDB for z/OS and OS/390 > Preparing Application Data for Upgrade >

Resolving Duplicate Row IDs

Upgrades: Release 6.x only.

Environments: All environments.

This topic is part of an upgrade process. See How to Perform the Upgrade.

During the upgrade from Release 6.x to 7.7, ROW_IDs from S_EMPLOYEE, S_CONTACT, S_ORG_INT, S_ORG_EXT, S_POSTN, and S_ORG_GROUP are inserted into S_PARTY.ROW_ID. Although ROW_IDs are typically unique across the entire schema, there may be rare instances when ROW_IDs are shared across these source tables.

If duplicate ROW_IDs are not resolved before the upgrade, the upgraded S_PARTY data will be defective because there will be mismatches between the base S_PARTY record and the corresponding extension table record.

After you install the Siebel Database Server software, but before you upgrade, you must identify and resolve any duplicate ROW_IDs in your Siebel Database.

To resolve duplicate ROW_IDs across source tables

1. Run the script, Find_DUP_S_PARTY_ROW_IDs.sql, located in the following directory:

   WINDOWS: DBSRVR_ROOT\DATABASE_PLATFORM

   UNIX: DBSRVR_ROOT/DATABASE_PLATFORM

   This SQL script generates a list of duplicate ROW_IDs (if any) and the tables containing the rows.

2. Choose one of the records that has a duplicate ROW_ID.

   Choose the record with the smallest number of references to other records.

3. Copy the record.

   This creates a new record with a unique ROW_ID.

4. Use the Merge Records command to merge the original record with the copied record.

   Records that were associated with the original record are now associated with the copied record.