Archives of the TeradataForum

Message Posted: Wed, 15 Nov 2006 @ 15:44:55 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hi David:

I haven't worked specifically with UTF data, but from your output showing character set, record and everything - it looks correct. Not that this is the fix, but have you tried changing company name by 4 characters so that the lengths match? If this is a UTF issue, I would think that records containing UTF data might load and those without would fail due to length. More of a curiosity question than anything else.

Since it is looking for UTF-8 and failing, is it possible that Oracle is not writing the data in UTF-8? Have you been able to look at the actual hex for a character and verify that it is what it is suppose to be?

Otherwise, have you considered using OLE DB via the ASXMOD command? It would go directly to Oracle and eliminate the need for an intermediate file. That have been many postings here regarding this technique and general comments also indicate good performance. Any thoughts from those of you have used the AXSMOD to Oracle?

Just a thought and sorry I don't have any words of wisdom based on previous experience with your situation.

Regards,

Michael Larkins
Certified Teradata Master
Certified Teradata SQL Instructor

	<Prev	Next>		<<First	<Prev	Next>	Last>>