DSXchange

We saw problems with IPC stages whenever the column metadata (datatype, length, display) defined in the stage didn't match what was coming from the source. Correcting the metadata helped but never resolved all of our problems. In the end, we removed the IPC stages whenever a job would fail with this...

If you're using Oracle and want to use SQL, I would consider regexp_replace. SELECT regexp_replace(my_text, '\D', ''), regexp_replace(my_text, '\d', '') FROM ( SELECT 'Th4is298 is a 12 test' AS my_text FROM dual ) REGEXP_REPLACE(MY_TEXT,'\D','') REGEXP_REPLACE(MY_TEXT,'\D','') ----------------------...

Are you compiling the job from a Windows Terminal Server client? We have seen issues with job sequences when complied from TS.

Thanks for the info chulett. I had suspected that the hashed files were being created as something other than Type 30, but I've never used anything else in five years, and figuring it out took a back seat to recovering.

I know this is an incredibly old thread, but I was hit by the same issue today. I thougt I would share the solution. When bringing up the Debug Window, there is no information showing in either of the two sections that should show the various columns and their values. The Local Name, Local Data, Wat...

Just so it's clear, setting the Transaction Size to 1 will give you what you want, but can also severly reduce the speed of your job.

Have you tried setting the Array Size to 1 on your target stage?

It's worth a lot. I keep telling people the best practice is whatever works today.

There's no such thing as a "sparse" lookup in a Server job, you mean a straight reference lookup. They're similar, sure, but that nomenclature is specific to the PX product. Good point. I guess I've been hanging around with my Enterprise Edition cohorts too much. I was able to get the DRS...

Well, I've spoken to our SQL Server DBA without any luck.

I also tried pulling the data from the source into a Hashed file for the lookup rather than the sparse lookup without any success.

Can anyone provide some help? Thanks.

Does anyone know exactly how long is "too long" for a SQL query in the OCI stage? I'll ask IBM, but hoping for a speedier answer here, if anyone knows. Trying to cut one down from over 2000 lines, ~40k bytes. Thanks. That's good info to know. Also, I'm impressed that you have a query over...

We use use both stage types. Performance always seems to be the same.

If you just have to have a favorite, try color. I prefer red over green.