DSXchange

There's a parameter in the Sort stage for temporary files and it defaults to "/tmp" I believe. Any chance you are running out of space there?

Which plugin? Try them one at a time to determine which one is having the issue. Could be both... Also ensure that you've done all of the setup work required by each stage - this is all covered in the Installation Guide and/or the readme that comes with the Server. For example, the OCI stages need $...

Did anyone go in and change the 'defaults' via the Director? This has been known to case the problem you are seeing, even if you don't touch the password field.

Welcome aboard!

Sorry that your first post generates this as an answer, but... no, can't be done.

Do all of your 80 files have the same metadata and the only difference is the delimiter? Seems like at best case you'd need 2 jobs to do the processing, at worst case - 80.

This is the user defined SQL UPDATE Dealer SET DEALER_DESC=:3 WHERE DEALER_CODE=:2; INSERT INTO Dealer (DEALER_KEY,DEALER_CODE,DEALER_DESC) VALUES (:1,:2,:3) There's your problem - all variables must be bound, so you can't get away with not mentioning :1 in your first SQL statement. And just to fur...

That said, someone (Craig??) mentioned a Windows script to do backups. posted either here or on Ascential DeveloperNet. FWIW, there is a nice little batch file posted in the File Library on ADN to automatically do a client-side export of all projects from a server. Combined with a registered versio...

How about your Operating System? UNIX can 'merge' files based on keys, from what I recall. You may be able to do something before job to get it ready for loading, perhaps even using the 'Filter' capability of the Sequential File stage instead of in a before job routine. At worst case, use the Merge ...

We don't let DS actually do the bulk load, only generate the data and control files for us. We then do the actual loading using korn shell scripts to parallelize the loading. That being said... When loaded via DataStage, does it generate the 'normal' discard or bad files if there are errors? You may...

It's here Ray. Scroll down to near the end of the thread.

This is a pretty straight-forward thing to accomplish. First, pull existing key column(s) from the target table into a Hash file. Then, when preparing data for the target table, check the hash file for the current key and split off two links from the transformer based on the result of the lookup. On...

And did you change which one was marked as a 'key'?

Supposed to be coming in the next (Trinity) release.

Hmm... doesn't sound like it's the quickest/easiest way yet.

What's wrong with something more 'normal' - "yyyymmddhhmnss" for example. Still a numeric sort with time down to the seconds and a wee bit easier to manage.

Interesting, it didn't take any deep sea diving at all. Ascential Support suggested I open the problem job (which I could after acknowledging the error 15 or 16 times) and then save it to a new name. Taking the newly created RT_CONFIGxxx hash and copying it over the top of the corrupted one fixed th...

I know the name of the job and was able to successfully export it. However, the corruption of the RT_CONFIG hash makes it impossible to either delete or import it (as the first thing import tries to do is delete it) - all I get is the stupid "Cannot open" message over and over again. I think this is...