Home Page for the TeradataForum

Archives of the TeradataForum

Message Posted: Tue, 29 Apr 2003 @ 16:27:14 GMT

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

Subj:   Re: Collect statistics..
From:   Anomy Anom

<-- Anonymously Posted: Tuesday, April 29, 2003 11:18 -->


It seems to me that the general responses to this question are all quite good. Please will the purists ignore this message!!!!

However at the risk of being very controversial, it can be at times better not to collect stats if you have a tight batch window.

If your data is very volatile, your stats will become stale quickly and in this case, you are better off not to have no stats than have stale ones. In this case, the random amp sample will be better than using old stats.

I do appreciate that this may lead to inconsistent query performance.

I would recommend, however, that there are stats on at least the primary index, if possible.

Finally, fresh(ish) stats are really the best option and the degree on which the data changes should determine what strategy you can use. Maybe it is good enough to collect at known quiet times (e.g. weekends, public holidays etc). The quiet times will obviously depend on your business cycles.

Hope this helps

  <Prev Next>   <<First <Prev Next> Last>>  
  Top Home Privacy Feedback  
Copyright for the TeradataForum (TDATA-L), Manta BlueSky    
Copyright 2016 - All Rights Reserved    
Last Modified: 27 Dec 2016