DSXchange

mandyli wrote:Any performance improvemnt .

Allegedly (from IBM).

The biggest difference is that Connectors can be used by other products in the Information Server suite; Enterprise stages are local to DataStage.

I don't believe that's true, but don't have time to check right now. Are you talking about a File Set stage or a Lookup File Set stage here?

At first glance it appears that you are using a Sort/Merge collector (on the input to the Teradata stage?) and have wildly different row counts coming from the different partitions. You claim to be using Round Robin, but there's an APT_ParallelSortMergeOperator featuring prominently in the error mes...

What else is happening on the LAN where the client machines are?

Once upon a time I encountered a day when the network was woefully slow, having been OK the day before. It turned out that 75 people were watching a major sporting event live in their browsers. Vale bandwidth!

Or maybe you want to be pedantic and recompile Transformer stages whether or not it's needed. Compiling Transformer stages takes considerably more resources than compiling most other stage types.

If you are using an ODBC stage, then it can accept and properly convert a DataStage internal format date. If you are using an Oracle stage, then you need to present a string in correct timestamp format ("correct" = matching the Oracle DATE picture).

Did you check the paths for the user whose ID is used to run the jobs?

Probably. But I don't know how. I'll defer to the XML experts on this one.

Are the data PRECISELY the same in both environments? What's common to the records that are being dropped? (Direct them to an extra output from the Filter stage.)

Have MJC generate a report and tell us, from that report, whether the routine was picked up and whether it was compiled.

I just did! Only run another instance when there are eligible candidates to run and there are fewer than four jobs already running.

Cold mornings but not bad otherwise.

Welcome to the leading edge!!! (My project will be getting there in a month or so.) IA Projects aren't actually created on disk anywhere. When IA is installed the DataStage project ANALYZERPROJECT (unless you rename it during the installation) is created. IA writes osh scripts directly - it runs Dat...

The requirement does not specify multiple-character delimiters.

Only one possible reason - the Perl script specifies to do so.

OK, there is a second possibility, that the special characters are already in the DSX file being processed.