Archives of the TeradataForum

Message Posted: Sun, 01 Nov 2015 @ 07:38:35 GMT

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

It is more important to collect statistics on non-index columns that are used regularly in joins, than the index columns. If your data is distributed well, then the index column statistics obtained by random AMP sampling should suffice, if data is skewed then keep these statistics also.

Cheers

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