Jobs aborted in the middle of dataloading into warehouse

[sanjay]

Hi All

I do have the 10,000 rows of data and the job get aborted while loading into warehouse after loading 5000 rows. How to hadle this situation to load it from 5001 record In Datastage PX not in server


Thanks
Sanjay

[gateleys]

Hi All

I do have the 10,000 rows of data and the job get aborted while loading into warehouse after loading 5000 rows. How to hadle this situation to load it from 5001 record In Datastage PX not in server

Thanks
Sanjay

First, find as to why the job is getting aborted. Fix that. How are you handling your rejects? Second, make sure that the 5000 rows that were passed from your job have been commited to the target, so that you can restart from 5001 row. If the rows per transaction was not set to facilitate this, you will find that the rows prior to the abort will have rolled back. You might have to do a rerun of the entire job.

gateleys

[gateleys]

Ooops!! This a post for a parallel job. I would not know how that would work. But the database principles should remain the same.

gateleys

[vcannadevula]

Moree information neede ,could you post the logs???? That might help in resolving the issue

[kwwilliams]

How to hadle this situation to load it from 5001 record In Datastage PX not in server

Why? Why would you not just send the rows to the database again? Set your database stage to insert then update. If your tables has a primary key or unique index on it, then the insert will fail due to a constraint violation and it will then update your data with the same values that you currently have in your table. No harm no foul.

Now if you don't have a unique index, you would update then insert. But I would strongly recommend your table have a unique index.

[ray.wurlod]

Did the 5000 rows actually make it into the target table? Or did the entire transaction roll back? Do you use a staging area of any kind between your ETL and your actual load process?

[kumar_s]

First try to load into a sequential file and make sure that you dont have any constraints set in your jobs. Like 'Abort after' or 'Read first n'...
If no, check the database if you can able to load the same set of data manually.
Make sure after 5000 records, it doesnt violate any referntial constrains or uniqness...