DSXchange

Why use a Routine Activity stage to do that when there is a specific Job Activity stage available? If you lay it out like T42 showed, running the three links into a Sequencer set to 'All' with your other jobs downstream from there, the three jobs will run simultaneously and will only move onto the f...

It generally means you are processing more rows than the stage can handle. It would help if you can presort your data on the grouped fields and then assert that in the stage by marking those columns as sorted.

Is that a possibility?

You might want to post the actual 'unlock' process you were given... what 'Universe commands' you are executing.

Post the entire text of your routine, please. Use the 'code' tags to wrap it so that it is readable... you can verify that by using the 'Preview'. Without those tags it left justifies everything and is hard as heck to read.

I wonder if there is really no possibility for that kind of controll in DataStage. What about an external scheduler? e.g. Tivoli Has someone a hint? Sure, and it's been mentioned. Hand code. The ones that I know of - Peter has written one, Kenneth has, even I have on a smaller scale. I suppose you ...

Oh, duh. Sybase IQ - sorry I didn't notice that bit.

we generated a million rows of data and tried to load on a small AIX server in our office here. Same result, 1,400 rows per second even when loaded directly via oracle loader.... When you say 'loaded directly via Oracle loader', you mean sqlldr and not the Ora Bulk stage in automatic mode, yes? And...

Not *quite* sure what the issue is here... By 'parameter' do you mean bind variable? And I'm pretty sure that an expression where you compare something (like a parameter) to something in quotes forces a string comparision and an implicit cast. I haven't had issues like this using the Sybase OC that ...

Oh... gotcha! I didn't quite follow what you meant by 'ten parallel streams' for some reason. Cool.

You could always write a server job to do it... source from the hash and write out to a sequential file stage.

True. Shouldn't be any problem loading the entire hash file into memory.

Hmmm... pretty sure that's what I said.

Put the first file's single record into a hash file with a hard-coded key value - like a '1'. Then stream in your second file, do the lookup and add the column to the output.

Easy Peasy.

paule wrote:It would seem that the job status would show the final status of the job after the second job failed.

Sure sounds like it should work that way. Perhaps you could open a support case with Ascential and see if there is a fix available.

The 'old fashioned' way to do it is to run an Otherwise trigger from each job to a single Sequencer stage set to 'Any'. From there you can link to a Notification Activity stage and it will include the status information from all jobs that have run up to that point, including the failed job.