DSXchange

Hi All,

In Server jobs, we can take the rejected records by using the reject row option in transformer or using O/pLink.REJECTEDCODE<>0.

My job flow is: Seq file---->Transformer---->Db2 Database

Simillarly in 8x, how can we take the rejected records from the target. We have a option called Otherwise/log in transformer but that would reject only if we have any constraints to the o/p link.

So as of now am using Basic transformer in place of Transformer to get the Target rejected records.

Could anyone suggest some better solution to this.

Thanks,

Senthil P

Which target stage are you using? DB2 Enterprise or DB2 API? What load method? Never worked on DB2 but oracle stage has an option to reject data when used as a target. And DB2 should be no different. The documentation says DB2 Enterprise and DB2 API support reject links.

This is not a Version 8 issue, it is a Server vs. Parallel issue. In Parallel jobs, the rejects are initiated in the stage that causes them - so you need to enable rejects in the DB2 stage and make a reject link from there.

I wrote a blog post and a wiki page about it some time ago that shows how to avoid row leakage in parallel jobs.

Hi

First you have take any reject target stage ( Dataset, Seq, OracleEE, DB2)
Linked the Transformer and Target stage with Linktab.
double click on Linktab there you have option enable reject link.

The regular LinkTab is ________________ like this.
Reject Link Tab is ------------------------ like this.



you can do it....

senthilt1 wrote:Hi All,

In Server jobs, we can take the rejected records by using the reject row option in transformer or using O/pLink.REJECTEDCODE<>0.

My job flow is: Seq file---->Transformer---->Db2 Database

Simillarly in 8x, how can we take the rejected records from the target. We have a option called Otherwise/log in transformer but that would reject only if we have any constraints to the o/p link.

So as of now am using Basic transformer in place of Transformer to get the Target rejected records.

Could anyone suggest some better solution to this.

Thanks,

Senthil P

The tricky part is knowing about and using the invisible columns that get generated down a database reject link in a parallel job. You can see them if you send your reject output to a Peek stage and turn column propogation on.