DSXchange

rohitagarwal15 wrote:we are not using any NLS parameter also in the job.

Did not mean in your job, I specifically meant the NLS default settings in both databases. One example would be NLS_TIMESTAMP_TZ_FORMAT which you can check with your DBA.

Your syntax was fine. What was the actual resolution?

I think you were clear... I for one have no idea. It doesn't sound like expected behaviour, what exact 8.x version are you on? Current on your fixpacks? If so, I'd get this in front of your official support provider if you haven't already.

I don't know SQL Server but it looks like you may be able to select a GUID via a Lookup or even as a source:

SELECT NEWID()

or perhaps

SELECT NEWSEQUENTIALID()

If that works, you could then pass it to both targets.

And yet Windows itself is not case sensitive. Obviously, some aspect of this was however.

I don't think you'd need to "hardcode" any transformation logic. I've had to do something similar in the past and put together a generic routine that would count the number of delimiters and then use Field() to take the count+1 delimited field from the full path.

From what I recall, the DOS option would appear when you specify a delimiter string rather than a delimiting character.

Can you verify the points in my post? I'm assuming Oracle is doing this, not DataStage.

For both questions, I'd really like to see the FD or whatever metadata you have on the incoming file. And that DecimalToString() conversion should have worked just fine if your incoming data is in fact a decimal, something you really didn't clarify or confirm. Any details you can provide on exactly ...

Yah, wondering if this is Oracle issue with a TIMESTAMP WITH TIMEZONE field. If so, I'd double-check that the NLS defaults are setup the same, source to target.

Those would be typical approaches in the Server world.

Confirm they are being used - have a DBA monitor the load while it runs and check the explain plan.

The constraint idea could work as well, I would imagine.

Downstream transformer. If that function is not available in a PX transformer, then perhaps a Server job or a Server Shared Container to leverage a BASIC transformer.

No, not that I recall. And the Server docs mention that the delimiter is the character that delimits the field.