DSXchange

If you can not View Data correctly in the job that reads from Oracle and populates data into the hashed file, then the mapping problem is in that job. You need to investigate. It may be that you need RAWTOHEX functions in the derivations for some of the Oracle columns.

In the Lookup stage you need to open the rules editor and change your "lookup failed" rule to output to a reject link. Then you need to provide a second output link (it will be painted with long-dashed lines) to capture those input rows for which the lookup failed. On that stream of proces...

I think tshort is the toothless version of the tsort operator. DataStage is inserting tsort operators (and alerting you to the fact that it's doing so) because sorting is required on the input links to the Join stage and your design does not provide explicit sorting on those links (so DataStage assu...

I don't believe IBM's "tool" has a name. It is only available through their services organization, and would be sold to you as a "migration service" of some kind with the specifics laid out in a Statement of Work.

I tried to start my car this morning.

But it didn't work out.

Please suggest.

Your issue is no easier to solve BECAUSE you have not provided any useful diagnostic information. What error message did you get?

Do you even understand what the PATH and LD_LIBRARY_PATH environment variables do?

Perhaps there is nothing that matches Jobs/*/*.* in your Repository.

What happens if you try Jobs/*/* ?

You need TWO outputs from the Transfomer stage.

You need to write TWO files.

You can put them together afterwards.

What ever happened to :1 ?

Your problem has nothing whatsoever to do with buffer sizes. It has something to do with writing to a job log.

Yes, that's how I answered.

(Actually I think the blocking operation in the server shared container might not be permitted. But everything else I suggested, including the use of server shared container, is perfectly legal for parallel jobs.

You want to sort odd numbered rows and even numbered rows separately? Then separate them into two streams and sort them separately. Use a Funnel when re-assembling the rows into a single stream.

The first requirement is that you have an IPC stage in your design. I don't think you do.

That question is too vague. One way would be to construct a server shared container containing a UniVerse stage or Hashed File stage and use the intrinsic dynamic normalization capability of that stage type. Another way would be to use the Complex Flat File stage. This has the ability to handle OCCU...

prasad.bodduluri wrote:you can increase ipc stge buffer size try once, i think it will solve your problem

Why do you think that changing buffer size on an IPC stage will affect log writing behaviour?

You don't. DSD.WriteLog is an internal routine (this is indicated by its "DSD" prefix). You look at the arguments you gave to whatever it was that you used to write an entry into the log (for example DSLogInfo() in a routine), or you involve your official support provider because there may...