Page 2 of 2

Posted: Tue Nov 08, 2005 9:34 am
by DSguru2B
hey,
guess what i found....
when i use the updte action as "Delete existing rows only". the job executes successfully. but any other update action forces the job to abort horribly.
why is that?

Posted: Tue Nov 08, 2005 11:17 am
by DSguru2B
Guyz,
The problem lied in metadata mismatch. Thanks for your help Kcbland. You sure directed me in the right direction. From the sql server, the key was set to bigint. when i changed that to integer within datastage, the job executed fine.
The target table still has bigint, but it executed fine.
Now thats another question huh?
No seriously, thanks once again.

Posted: Tue Nov 08, 2005 11:17 am
by DSguru2B
Guyz,
The problem lied in metadata mismatch. Thanks for your help Kcbland. You sure directed me in the right direction. From the sql server, the key was set to bigint. when i changed that to integer within datastage, the job executed fine.
The target table still has bigint, but it executed fine.
Now thats another question huh?
No seriously, thanks once again.

Posted: Tue Nov 08, 2005 11:26 am
by kcbland
Hence the bind issue, trying to stuff one datatype into a non-supported value for the driver. Well, sometimes the last shot is the one that works. We knew the SQL was the issue, just the datatypes when the connection is made and the SQL prepared was the issue.

Don't worry about the datatype inconsistencies between your job and the actual table. Just put an Annotation on the job so that metadata tweak isn't lost or forgotten if the metadata is ever reloaded.

Posted: Tue Nov 08, 2005 12:35 pm
by DSguru2B
I will do that. Thanks once again.