DSXchange

There's probably more to it than that, if it works anything like sqlplus. It will begin and exit just fine as nothing went wrong with the session itself and could care less that everything you ran in the session blowed up. You may need to take extra steps to detect those problems, as I mentioned in ...

You need to incorporate some mechanism in your script to capture those BTEQ errors and propogate them back up to DataStage. I've had the same issue in the past with sqlplus, all the script (and thus DataStage) knew what that it successfully launched the sqlplus session. It had no clue that the sql i...

Yes, what database is this coming from? All these examples and pictures and nary a mention of what database you're working with. May be useful information after all.

The key information seems to be this: Graphics environment not configured correctly. Refer to Release Notes for instructions to resolve this problem. Have you referred to the release notes yet? Seems like one of those kind of things related to where you may be running this - may need to be from the ...

You could also use the DS.TOOLS menu for that.

Interesting... me, I would just tee that off to a sequential file.

What purpose was the hashed file serving in the job?

Yes, they can be 'problematic' at times with how they handle nulls in key columns, but other than that no.

Not only not really what was wanted but wouldn't actually work either. While you can drop/exlcude columns downstream from the reader, unlike a relational source you cannot pick and choose your columns.

Yes, outside or inside DataStage is perfectly feasible. Check out the "dsjob -jobinfo" option from the command line or the DSGet* API function equivalent.

OK, but the conversation is specifically around the compatability of that XML format with the other import methods to see if using istool to get it back in is one's only option.

Hopefully someone can test and verify this.

When does this 'abort' need to happen? Any check like that can be made well before it ever hits a DataStage job, if that's what you need. Or you could count it first, pass that count to the job and have it validate that count as part of its processing... usually involving an Aggregator doing the cou...

The short answer is 'no', at least not directly. There may be some 'workarounds' for this but I'll leave that for others to post.

Of course, you need to check everything that could change from environment to environment. That would include the config file used, actual data, source/target metadata, etc etc.

I've personally seen much more evidence of 'forwards' compatability rather than backwards, but of course there is. However, it's not all all inclusive matrix. Found this online, however as with the poster of this information, I have no MetaLink access either so cannot confirm: According to the guys ...