DSXchange

btw, this odd behavior only occurs on RESTART. Steve - can you clarify exactly what you mean here? Do you specifically mean only with a Sequence job that is in an 'Aborted/Restartable' status? Or any job with that status? Something else? I'm also curious about the -local option you are using. I've ...

You need to presort your data to support your grouping keys and assert the sorted order in the Aggregator stage.

Have you looked to see what's in your ScreenOutput variable?

What's your volume? As noted, I'lll wager you need to sort your data and then assert that in the stage.

What result are you getting?

You don't need the 'sh' part on the front, simply put the fully pathed script in the Command box and then any parameters / arguments (if needed) in the Parameter box.

Run the job and then post whatever messages in the log relate to that stage.

Perhaps you could verify that, in fact, the indexes you believe are in place actually are. I don't recall the syntax you'd need off the top of my head, however. I say that because the only time I see rows/sec run at '<1' in this situation is when there are no indexes in place. And yes Ray, we all kn...

Answer seems pretty obvious. The DRS stage does not support numbered bind variables. You need to stick with the positional (yuck) question mark variety. Besides, I thought only Oracle used the numbered ones? As to your second question, carefully read the post you linked to. The question about 'dynam...

What have you tried so far?

Thanks, good to know.

We'll see if that changes anything, Ernie. I believe the error lies elsewhere. I'm curious though, does the Sequential file stage just make that specification file more obvious because it aborts the job while the XML Output issue could be more of the 'Silent But Deadly' variety? Or are you advocatin...

It's all about your version of DataStage. You must have 7.5x2 or 8.x in order to run PX jobs on Windows. With others you can build / compile but as noted they cannot run there. No patch or O/S upgrade will change that.

No, the technique is perfectly valid. You build the hashed file with X keys and lookup on < X in the UV stage. The reason for the difference will be the simple fact that you must not be doing the same thing in the Hashed File / UV solution that you were in the ODBC one. Without all of the gory detai...

Why are these two dates in a different format than the rest?

CONTRACT_EFF_DATE = 03/01/2007 CONTRACT_TERM_DATE = 12/31/9999

Consistent handling of dates is key to success.

Just keep in mind the fact that it isn't a "supported platform" for the client. Not saying it won't work just that it may be difficult to get help with it if you have problems.

i have a table in the oracle database. i want to do a transformation on one column of that table and load back into the same table with the same job. that means i have to give the same table name in the oracle stage properties for the i/p and o/p. I want to know whether it works or not? I understan...