Archives of the TeradataForum

Message Posted: Thu, 28 Feb 2008 @ 09:57:48 GMT

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

Omer,

If you have time, try to stage the following experiment.

Make a copy of your EVENT table but change its definition to MULTISET and turn the UPI into a NUPI (drop the uniqueness constraint). The data should be the same as in the original EVENT table.

Run the same ML job on the same file into this new table.

Report the run time.

This should give us and you an idea of how much time your system spends on uniqueness checks and then we can go from there.

Victor

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