Archives of the TeradataForum

Message Posted: Thu, 03 Jul 2003 @ 14:29:14 GMT

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

Here're three things I'd look for:

1- Make sure your target table is empty... This'll reduce the journaling overhead

2- Make sure your target table has the same primary index as the source table, otherwise, you'll redistribute the entire source table.

3- Check the uniqueness of the primary index on the target table. If you have a large number of PI occurrences with more than 10000+ counts, you could be spending all of your time validating uniqueness of rows in the Non-Unique Primary Index (NUPI) chains. You can get around this by defining your target table to be multi-set (this bypasses the uniqueness check, but this might be ok if you're copying data from an existing table.).

Pat Belle
Certified Master

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