Siebel Database Upgrade Guide > Preparing Siebel Application Data for Upgrade >
Preparing Siebel Customized Seed Data for Upgrade
Upgrades from: All Supported Siebel releases.
Environments: Development, production test, production.
Platforms: Windows, UNIX, IBM z/OS.
Modified seed data are not upgraded, even if the upgraded seed data are required for an application to run normally, with the following exceptions:
- If you modify seed data through the GUI or through EIM, the repository merge preserves the modifications using the same logic as for other repository objects.
- As of Siebel CRM 8.0, customizations to seeded workflows in the Prior Customer Repository are merged into workflows in the New Customer Repository by the repository merge process.
Before starting the upgrade, evaluate whether you need to retain your seed data modifications. If practical, consider discarding the modifications. After the upgrade, you can reimplement those modifications that do not interfere with operation of applications.
Seed data are records provided in the Siebel Database tables as part of a Release. Seed data provides information needed to run a Siebel application. Examples of seed data are the values in a List of Values (LOV) definition, default mappings of views to responsibilities, and predefined queries.
The ROW_ID value for seed data records provided in a Release begins with 0 (zero). In addition, these records have a default LAST_UPD value of 1980_01_01. The LAST_UPD field stores the date when the record was last updated.
If you modify a seed data record, the value of LAST_UPD is changed from 1980_01_01 to the date the modification was made. If you add a new seed data record, the ROW_ID does not begin with 0, and the LAST_UPD value is the record creation date.
When you perform the upgrep part of the upgrade, the dataimp utility upgrades seed data records as follows:
- For records where ROW_ID begins with 0, it erases seed data records where the value for LAST_UPD is 1980_01_01, unless prevented by scripting.
- Dataimp replaces these records with those contained in seed data files included in the Release. In the new records, the value for LAST_UPD is 1980_01_01
- Dataimp does not erase and replace records where the value for LAST_UPD is later than 1980_01_01. Instead, for each of these records, dataimp writes an error to the log file. The error is benign and does not cause the upgrade to fail.
Use these error entries to verify which seed data records were not changed.
For seed data records you have modified, you can discard the modifications by changing the value of LAST_UPD to 1980-01-01. This causes dataimp to replace these records with those from the new Release. For seed data records you have created, the upgrade process retains these records.
To discard seed data modifications
- In Siebel Tools, select the Table object.
- In the list applet, query for "*" in the Seed Filter column.
The query returns a list of tables containing seed data.
- Use one of the following scripts to set LAST_UPD to 1980-01-01 for customized seed data records in these tables. In the scripts, tablename is the name of table containing seed data.
UPDATE tablename SET LAST_UPD = TO_DATE('1980-01-01',
'YYYY-MM-DD') WHERE ROW_ID LIKE '0%' AND LAST_UPD > TO_DATE('1980-01-01', 'YYYY-MM-DD')
IBM DB2UDB and IBM DB2 for z/OS
UPDATE tablename SET LAST_UPD = TIMESTAMP('1980-01-01-00.00.00') WHERE ROW_ID LIKE '0%' AND LAST_UPD > TIMESTAMP('1980-01-01-00.00.00')
Microsoft MS SQL
UPDATE tablename SET LAST_UPD = CONVERT(DATETIME,
'1980-01-01') WHERE ROW_ID LIKE '0%' AND LAST_UPD >