DSXchange

This will happen if you have more than one instance of DataStage installed on your QA server.

In a PX transform stage you would do:

What are your column definitions for the source dfloat and the target decimal?

Are you running a 1-node configuration (if not, try it and see if your 'jumbling' might be coming from your repartitioning)

sarangadhar wrote:...-d 06/11/2006...

Hmmm, could it be an expired license?

Hello sarangadhar and welcome to DSXChange.

Are you running this on Windows with a userid that is a local administrator?

see this reply to the same question in the Server forum

I am not sure what you mean by "root A" and "root B" on the same server. You can install one or more DataStage instances on any given machine (with 'machine' defined as an instance of an Operating System). The software will allow you to do this using the same license keys. Your contract with IBM doe...

Use
It is more efficient as well.

It would be expected if you defined the column as CHAR(5) insteado f VARCHAR(5).

The mathematics of a CRC32 computation are bit-oriented so should not be different unless you are using different code. Are you certain your versions of DataStage are exactly identical?

DS/Basic arrays can be dimensioned more than 32K. The code references "$mam" - what is that value at runtime?
Perhaps the array size referenced here is for your transfer buffer to/from the database.

Hello and welcome to DSXChange. Please post the complete error message from your log, or at least that part which details what led up to the SIGSEGV. Disabling the combination will help locate which source object might be causing the problem; best leave it to TRUE in order to make debugging easier.

Just fix and address one problem at a time. Ignore Oracle for the moment, just output to a peek stage or a test flat file.

If you output the timestamp column with your StringToTimeStamp() function does it show a correct timestamp?

He is writing a PX job, so OCONV() won't work.