'Enable partitioned reads' at Oracle Connector

sohasaid · Post by **sohasaid** » Tue Oct 16, 2012 5:43 am

Hi,

I'm trying to read data from Oracle using parallel mode, my table is not partitioned so I'm using 'Rowid Range' option after enabling the partition reads features at Oracle connector stage.

The job was working fine at Dev server and reads were in partition mode, but after migrating to Prod server I experience a weird behavior: the generated select statement from oracle connector stage is WRONG.

It puts 'where' clause before 'from' and it passes some columns names in a wrong way!! As follows:

Code: Select all

SELECT URL, FAX_NUMBER,CUSTOMER_ACCOUNT_ WHERE  (MOD(DBMS_ROWID.ROWID_ROW_NUMBER(STG_ACCOUNT.ROWID), 2) = 1) GROUP FROM STG_ACCOUNT.

Of course the job aborts because of the wrong select statement.

Any idea why it's not working on Prod while it's working before?

Thanks

PaulVL · Post by **PaulVL** » Tue Oct 16, 2012 8:28 am

Things I'd look for:

Same patches on DataStage servers (Version.xml)?

Same Oracle tool release levels on your DataStage server (client code)?

Same Oracle database version and patches?

sohasaid · Post by **sohasaid** » Tue Oct 16, 2012 9:45 am

what exactly you want to look at?

DataStage client version: 8.5
Oracle version: 11

PaulVL · Post by **PaulVL** » Tue Oct 16, 2012 12:24 pm

"I" don't want to look at anything, since I'm pretty lazzy. I want YOU to look at your DEV environment and compare the answers to your PROD environment.

8.5 is not the answer I was looking for. Version.xml will tell you which patches (and in which order) they were applied to your system.

You need to investigate as much as possible why DEV is acting differently than PROD.

Do you recompile your code in PROD? (hope not)