Archives of the TeradataForum

Message Posted: Thu, 13 Aug 2004 @ 01:37:27 GMT

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

A subset. No, that won't fly. You have to cover the whole PI. And a secondary index on just the subset is not going to be useful to the optimizer unless there are also explicit terms (i.e. colref = constant) on some of those columns, or there are terms on the small table to make it very small.

If you have no way of limiting the number of rows in the large table, I think you'll probably have to take the redistribution hit.

Jon

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