Archives of the TeradataForum

Message Posted: Thu, 02 Apr 2009 @ 11:20:20 GMT

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

Hi Vishal,

Before Teradata can perform the comparison (the "Hash join" I seem to remember) it needs to convert at least one of the two character sets to the other. That takes quite some effort.

The instance I've encountered this phenomenon it was a mistake that one of the two tables was defined as Unicode. So it was just as easily fixed.

I cannot remember whether the Latin is converted to Unicode or the other way around (I would think the former) But to avoid the performance drop, I'd suggest you get the two tables in the same format. Perhaps a preceding job that converts the Latin table to a temporary Unicode variant?

Regards,

Menno van Tienhoven

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