DSXchange

I find that very reasonable. I could run the same job 40 times in a Sequence, but each occurrence would have to be associated with a unique stage name. If all it reported was the job name, there's no guarantee of uniqueness and I may not know which the heck one was the problem child.

The same way you should have been doing it all along - via the proper XPath Expressions in the Description fields of the output link. Import the metadata from an xsd and they will be created for you automatically.

The Multiple Job Compiler will compile routines. As to your other questions - no, yes.

But only to set it to its own value, basically a 'no op'. That's not a problem, per se.

Yes, it will. Still generally better to script something like that but you can do it that way if you like.

Okay... could have sworn that was the standard advice we've seen here in the past. Is this new with 8.x or has that always been the case? nm, just checked and even I have a wire driver for SQL Server here in our 7.x non-Enterprise version. Sheesh, I wonder what I was mixing this up with... or if I w...

Only if you've purchased UNIX ODBC drivers that support SQL Server, which is Windows only. Which is also why you don't have a stage for it.

No clue, I'll have to let someone who actually uses 8 answer that.

Yes, or the safer alternative: 'OK' for success, 'Otherwise' for failure.

If you mean you are getting inserts when you think you should be getting updates and thus creating 'duplicates', there would need to be a unique constraint / index over the columns you don't want duplicated for this 'combo action' to work. In other words, the first action has to fail for the second ...

Hard to do without seeing what your data looks like coming out of the sort stage, i.e the actual data your transformer / stage variables are processing.

Yup.

http://www.dsxchange.com/viewtopic.php?t=119271

Then I would dare to suggest that your logic was faulty. Captain.

pravin1581 wrote:is it not possible to use one stage variable in the derivation of the other one when it is created earlier than the latter one.

If that's a question, then yes - it is possible. They wouldn't be all that helpful if you couldn't refer to one in the derivation of another. Earlier or later.

Did you commit your deletes?