Page 1 of 1
Phantom Error
Posted: Mon Sep 01, 2003 6:00 am
by debajitp
I am getting this error while running my DS job. The data is loaded successfully but this warning is coming up just before the last statement, in the log. Pls help to remove this warning.
"DataStage Job 9 Phantom 2168
Program "JOB.660314562.DT.1248969383.TRANS1": Line 172, Variable "$R23" previously undefined. Empty string used.
Program "JOB.660314562.DT.1248969383.TRANS1": Line 172, Variable "$R23" previously undefined. Empty string used.
Program "JOB.660314562.DT.1248969383.TRANS1": Line 172, Variable "$R23" previously undefined. Empty string used.
Program "JOB.660314562.DT.1248969383.TRANS1": Line 172, Variable "$R24" previously undefined. Empty string used.
DataStage Phantom Finished"
Posted: Mon Sep 01, 2003 12:36 pm
by spracht
Paul,
the gurus all seem to have gone to holiday, so let me try [8D]: I had this kind of message for variables I was using within routines. To avoid the warning, I initialized them, e.g. $R23='' or $R23=0. Maybe you do a conditional initialisation, which is not executed for all of the records processed[?]
HTH,
Stephan
Posted: Mon Sep 01, 2003 12:47 pm
by kcbland
It would be helpful to describe your platform, your version of DataStage, and the components of your job design. Other than a transformer related error message, we have NO IDEA what your job is trying to do, and therefore can probably rattle of twenty different things to look at in order to resolve your issue.
My guess is shared container related, but not knowing your release or job design doesn't help.
Kenneth Bland
Posted: Mon Sep 01, 2003 8:21 pm
by kcbland
And yes, today is a US holiday.
Kenneth Bland
Posted: Tue Sep 02, 2003 1:17 am
by ray.wurlod
... and I've been in paid employment [:D]
Variables with names of this form ($nn) are the temporary variables that DataStage BASIC uses to store intermediate results in expression calculation.
This is occurring in a Transformer stage called TRANS1 in your job.
The source code can be viewed in a sub-directory called RT_BP9 for this particular job, and the actual expression that is causing grief will be at or near line number 172 in that code.
(If you're on UNIX, use set nu within vi; if you're on Windows, copy/paste the code into Word and switch on line numbers from File > Page Setup).
I say "at or near" because line number reporting is sometimes out by one or two when $INCLUDE directives are used to incorporate the contents of header files.
It suggests that somewhere, somehow, an expression is trying to deal with a variable that is in an unassigned state. A bit difficult to say more, without actually seeing the job design, or at least that part of it.
Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518