Issue with updating partitioned table
Posted: Wed Jan 25, 2012 12:41 am
My target is range partitioned table and partitioning key is month. My source is a sequential file which has around 15 rows which pass through a transformer before trying to update the oracle target( Oracle connector) based on key columns. I also have a reject link for the target with "Reject if row not updated" and "Reject for unique constraint violation checked".
When I run the job with the target link hash partitioned on update keys, it rejects some rows with "row not updated". However, for these rows there already exists a match in the target table.
However, if I hash partition the target link and check the in-link sort option, it doesn't reject any rows.
Can anybody please explain whether it is necessary to specify the sort option while doing a hash partition? Why are the rows rejecting if not sorted?
I am on DS8.5
When I run the job with the target link hash partitioned on update keys, it rejects some rows with "row not updated". However, for these rows there already exists a match in the target table.
However, if I hash partition the target link and check the in-link sort option, it doesn't reject any rows.
Can anybody please explain whether it is necessary to specify the sort option while doing a hash partition? Why are the rows rejecting if not sorted?
I am on DS8.5