Archives of the TeradataForum
Message Posted: Wed, 07 Feb 2008 @ 00:36:55 GMT
don't worry about case 2,3,4: They are less or equal 1% of the whole data.
For case 1:
This is 10% meaning 10% skew, which is also not "too" bad. For a FTS this means a 10% penalty vs. a perfectly distributed table. I don't know the size of your system, but 10% of ( 10M rows x 127 Bytes ) = 127MB "additional" work for 1 AMP does also look not too bad.
As long as the PI is the choice for many of your queries - stay with it.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|