Page 1 of 1

After job aborting or fail how to get next values

Posted: Thu Oct 16, 2008 10:10 pm
by deesh
Hi DS Guys,

When I run the job, target table has loaded 102 records after that due to some problems job has been failed or Aborted. After solving the problem I want the to load the data from 103 record onwards (I dont want to load from 1st record).

Posted: Fri Oct 17, 2008 12:45 am
by datisaq
I think you've to select the option at the target table as "Update and then Insert".This will update the existing rows and insert new rows...

DS gurus please correct me...

Posted: Fri Oct 17, 2008 12:49 am
by chulett
Short answer is "skip them". However - static source? Meaning when you rerun it, will the 103rd record in the second run be the same 103rd record from the first run?

Posted: Fri Oct 17, 2008 3:49 am
by deesh
How to skip the first 102 rows and moreover if I skip the rows, that job will can give good result. Please can provide the proper guidance in answer

Posted: Fri Oct 17, 2008 7:37 am
by jaybee223
Not a quick solution - you may try comparing the loaded records against the source and eliminate them. If upsert option does not work for you, have you tried using a transaction size of 0?

Posted: Fri Oct 17, 2008 8:28 am
by chulett
You "skip" them using a constraint in a transformer (@INROWNUM > X) but there's a lot more to it than that in the PX environment. And only you can say if it will "give good result" if you do that.

Posted: Sun Oct 19, 2008 10:29 am
by Nagaraj
Did this really work?

"constraint in a transformer (@INROWNUM > X)"" if YES then its really good one....!

Posted: Sun Oct 19, 2008 3:13 pm
by ray.wurlod
Yes/no/maybe. Remember that each Transformer stage instance (there are four if you are running in a four node configuration) keeps its own, separate, count.

Posted: Sun Oct 19, 2008 3:19 pm
by Nagaraj
Oh okay...thanks for the information.
in that case it wont be useful for me when restarting the job from the point of failure.