Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Tue, 12 Apr 2005 @ 23:52:46 GMT


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


Subj:   Re: Identifying Candidate Columns for Statistics
 
From:   Stover, Terry

If you're on R5, that's presumably the purpose of the stats wizard in the Teradata Analyst Pack. It's a 1.0 release though. With R5 you can also collect stats based on a sample instead of the full table.

On R4 I was downloading the accesslog sql statements into a sql server database and parsing them to count usage of the join and where columns. I had the advantage that the vast majority of our queries came from Business Objects, so they were very structured. SQL wasn't really the right tool for the job, but installing Visual Studio totally hosed my PC and I had to reimage.

You may want to consider tracking being more creative in your stats collection, some columns are more dynamic than others. In a prior life (on another db platform), I set it up to collect stats based on row counts changing >10% or quarterly. We collected stats on dates and order numbers every week since the values were always being added at the end of the distribution. Some columns like status codes or organization id's don't change very much.



     
  <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