DSXchange

Imed - best if you start a new post and let us know the details of your problem. Then we don't waste a lot of time with answers that don't apply.

I don't have any DB2 experience but in Oracle you can't simply select XMLType data from the database. They have a set of functions specifically for those kind of tasks. Not sure how equivalent the two are but if you have a specific XML data type rather than simply storing it in a CLOB then I'd suspe...

iq_etl wrote:Would you suggest just copying over the 8.7 jobs to 9.1 if we determine them to stay as server jobs?

Exactly... export from 8.7, import into 9.1 and you're ready to go.

We'd really rather not have to write conversions for all of these First thought? Don't. Just because you are moving to Parallel jobs doesn't mean every single Server job must be rewritten in Parallel. Use PX for new jobs where appropriate. Start the rewrite with those Server jobs that would most be...

Moved.

Probably. Be patient...

Also verify your data types and sizes, source to target. I for one do not expect that kind of warning going from a small field to a larger one, it is generally the other way round that causes issues.

Sorry but what are you attempting to do with the substring operators? The "[5,0]" part... remove it. Don't see any need for the "ceil" rounding either. What happens with a straight-up function call?

a) use the xml Stage instead. The xml stage requires an xsd, and it has a steep learning curve, significantly because it demands that you have familiarity with xsd's and are comfortable with hierarchical concepts. This would be the new stage that I mentioned earlier, Joyce. No input or output just....

You only use the INTACT keyword for partial schemas. Changing quotes to "double" is what fixed the problem in your subject line, removing the "intact" property is what let the fields be parsed properly. How did you end up with that option in the schema file?

You may need to include the drive letter if the files are on a different drive from where the DataStage jobs are executing. Otherwise a path is a path is a path and the Execute Command stage doesn't treat them as anything different from what the Sequential File stage would in a job.

If you are on Windows, change the 'echo' to a 'type'. Right now it's probably just echoing the string that is the name of your file rather than the contents. On UNIX you would use 'cat', on Windows the equivalent is 'type'.

Glad I could help.

Yah, we have a 'double-posting' issue here and they can even come several minutes apart depending on what the user does navigation-wise after the initial post. Usually I'm here cleaning them up but I was still asleep.

Time to involve your official support provider.

rahcomp1 wrote:One confusion here, In the technote lib file name is mentioned as 'libicui18n.so' but the file I have, has an extension '.a'.

This extension changes based on the operating system.

Of course.