DSXchange

My guess (like Arnd's) is that you have partitioned on a column (or columns) that have a skewed distribution (or maybe only five distinct values). This is something you need to check. If you want even distribution of rows, choose Round Robin partitioning (or Modulus, if it's a single integer sequenc...

If the data are in a sequential file, you have to read the sequential file (even to get the data into a persistent Data Set). So there's no "either/or" about it.

Investigate "multiple readers per node" property of the Sequential File stage.

Lock for update only sets a lock if the lookup fails (NOTFOUND is true). Yes, it does expect you to update the hashed file - isn't that the topic of this thread? When the job finishes any non-released locks will be released. (I suspect, but have not proven, that the lock is released when processing ...

du supports the h option in Dilbert, as in "duh" and "induhvidual"

Modularity is what you are achieving, and typically it is a Good Thing.

Not true in earlier versions, where the job status was actually output on stderr. I can't recall when this was "fixed", but it's somewhere in version 7.

Get some more knowledgable friends. The name of the file is DSParams, and the only job parameters stored there are those set up as environment variables in the Administrator client. Other parameters are properties of the job or shared container, and therefore are stored in the DS_JOBOBJECTS table in...

Are you still at RIL? I'm fairly certain there's already a routine there for doing just this task.

Please post those rows, so that others might check them. Also post the SQL statements from the reference input link and stream input link. Finally, advise what stage types you are using.

Or, if you use the one hashed file, disable read cache and write cache, and enable lock for update when reading.

narsi1111 wrote:yes i don't have pivot stage

Wrong. You DO have the Pivot stage. You may have not installed it, but it's there on your DataStage CD and can be installed in seconds using the Package Installer.

If you use DSGetJobInfo() from an after-job subroutine you can request the "interim status" - that is, the status with which the job will finish assuming it's not aborted by the after-job subroutine's ErrorCode argument.

IBM has no plans to offer a certification for server jobs.

Take the vendor's training class. This runs for four days. Rather more than be packed into responses here.

I am the DBA who support DataStage 7.52.
... my background of Datastage is zero.

These statements are mutually antagonistic.