1. Stop your Replicat on the Target side, most likely it already abended, goto step 2.
ggsci>stop replicat <name>
2. Comment out the MAP statement of that table in your Replicat parameter file
3. Restart your Replicat, so all other tables are still being replicated.
4. Use the LAG command to check the LAG of the Replicat, wait till the Replicat catches up.
ggsci>lag replicat <name>
How to tell it catches up? When the LAG result shows AT EOF
5. Stop both the target Replicat and source extract.
6. In source extract parameter file, add:
FETCHOPTIONS FETCHPKUPDATECOLS
N.B. FETCHOPTIONS FETCHPKUPDATECOLS is not valid for NSK Guardian extracts. Omit it.
7. restart the extract.
8. reload the table from the source to target using your initial load method after deleting existing rows in the target table.
9. Wait till initial load is finished, get the table back in your Replicat parameter file, put on handlecollisions parameter
10. Restart your Replicat
11. Wait till the Replicat catches up, stop the source Extract and target Replicat
12. Comment out handlecollisions in Replicat parameter file, and comment out “FETCHOPTIONS FETCHPKUPDATECOLS” in Extract parameter.
13. restart the Extract and Replicat.
This procedure applies to all the platforms, all databases.
