Home Page for the TeradataForum

Archives of the TeradataForum

Message Posted: Sun, 17 Jun 2001 @ 22:16:20 GMT

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

Subj:   Re: Why redistributing?
From:   John K. Wight

We all know that we should collect statistics on NUPI's. But do you collect statistics on UPI's as well? It's always been a contentious point but I have always don't them as a rule of thumb. Here's JK's general Collect Statistics 'guidelines' for what they are worth:

1. Collect on all UPI & NUPI columns (indexes)

2. Collect on all NUSI & USI columns (indexes)

3. Collect on all columns used in joins (PI's or not)

4. Collect on all columns used in qualifications/restrictions (i.e. used in the WHERE clause)

If you do this as you are initially designing (specially when you are doing a first out with canned reports (parameterized or not) and setting up your database - you will have a great majority covered. Also, when you are in initial testing/staging the systems for production. Access Logging on your development userid's is a very useful tool to capture SQL and see what the 'apps folks' are doing and gives you the chance to add collect stats on other columns.

Oh - we all know UPI's are unique, but does the optimizer know how many there are?

Just my thoughts for a sunny Sunday afternoon in 111F heat! (But it's a dry heat;-)

  <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: 15 Jun 2023