DSXchange

Thanks Ray ji. Its obvious, but I was'nt sure.

However I missed a point in my earlier post. I don't retrieve logs unless a job fails. However I have seen that even if job B aborts, seq job S returns a jobstatus=1. Isn't that misleading? Thats becuase you have "Automatically handle failures". Uncheck that, this way you will have atleast a warnin...

ray.wurlod wrote:Probably. Write it yourself in C, just to prove the point.

Nah. Even if there is a difference, it must be negligible.

Or, load a hashed file with your data and tweak the routine I posted to read that particular coclumn and retrieve the Median.

Thats a good idea, to retrieve logs for every job. But if you need for a the job that failed then as advised earlier, you need to come up with a way to, from the log, get the job name that failed.

pavanns wrote:i tried doing it with varchar all the way but does not work as it says cannot insert null into the column.

Is it possible that your getting null from source and trying to insert to a not null column?
None of the rows get inserted?

Craig, have any other bright ideas

Well, make sure while copy pasting you dont have a trailing space.

Well, you will have to have a mechanism to see which job fail. Probably traverse through the log that you create, get the jobname that failed and then get the log for that particular job.

Build a transform routine and not a before/after subroutine. FUNCTION myGetRowCount(JobName,StageName,LinkName) $INCLUDE DSINCLUDE JOBCONTROL.H hJob = DSAttachJob(JobName, DSJ.ERRFATAL) RowCount = DSGetLinkInfo(hJob, StageName, LinkName, DSJ.LINKROWCOUNT) ErrCode = DSDetachJo...

When I said "pops up", I actually meant, if any message gets logged in the log file. Its a figure of speech. Anyways.
Give us more details of your job design, are you using inter-process buffering or IPC stage etc. Need more info.

Thats odd. Did you try a re-install ?

Build another sequence with two job activities. The first job activity will call your present sequence and the second job activity will run JOB A. The trigger will be "Unconditional".

The job that is failing, is it loading to a table. You might have that particular column set as nullable at the datastage end but if its not nullable in the database, it will generate error.
As for the run times, it depends upon the data volume, if network is involved, then network traffic etc etc.

If you have to run them one by one then your good.