DSXchange

No, not really. I meant by tracking the number of rows processed, adjusted based on the commit interval, you can work out where you need to pick the processing back up from. This would rely on a 'static' data source. If you knew you needed to pick up processing from the 34,001 input record you bette...

No worries on that account.

Gotcha... thanks again.

We're quite happily using the Oracle Wire Driver against 10g databases with 7.5.1A - what version of DataStage are you running?

stivazzi wrote:- server job: i try all the ways (transaction size, handling, whatever you want); if i get an error (insert duplicate key), does not manage rollback!!! what has been inserted is inserted and that's all.

The job must Abort to issue a rollback.

Are you generating your own surrogate keys or the database is generating the keys? If the answer is yes then you could have the max(key) saved in a hashed file and then in the transformer, you could have the constraint in.key > max.key Or take a similar approach, but use @INROWNUM in the constraint...

No, he did not.

Is that ok to set on your pc? Give it a shot and see, let us know.

I know mine is.

ppalka wrote:Does anybody know how to set up ds to display properly national characters in ds view data?

As far as I know - you don't setup DataStage for this, you setup your pc. What NLS_LANG setting is your pc set to?

My Computer / Properties / Advanced / Environment Variables

Look under your Projects dir in the RT_BPxxx dir for the job number mentioned in the log.

ray.wurlod wrote:Whenever you like, provided you don't mind incurring the additional monitoring and maintenance overheads.

Ray, could you expand a little bit on this please? Not looking for a Full Wurlod or anything, but curious what you meant by 'additional monitoring and maintenance overheads'.

Thanks.

First off, why start a new thread? If you 'have a doubt' or want to ask a question that continues that coversation, do it in the original thread.

Secondly, what kind of additional 'information' are you looking for?

I guess that person would be me. I never said that 'dsjob' was a scheduler itself, but that you would be using 'dsjob' with whatever scheduler you ended up using. In other words, what you would schedule would be a scripted call to the dsjob command to run a particular job or job sequence for you, us...