Archives of the TeradataForum
Message Posted: Thu, 27 Mar 2008 @ 22:17:24 GMT
Basically I would say yes and so does the manual of Teradata. Anyway there has been a posting regarding statistics on certain columns where the performance decreased (or was it about indexing ??? I am not sure). So make sure when you do collect statistics that you take a look on the explain.
Anyway what really helped me to increase performance is using derived tables within your query to aggregate the records as soon as possible and before you join other tables together (makes sense if you have huge fact tables)
Eg. a is representing the fact table, b is the lookup table
Select b.field1, b.field2, sum(a.field3) as field3 from Table1 a Left outer join Table2 b On a.field1=b.field1 And b.field2=b.field2 Where a.field1=123 And a.field2=345 Group by 1,2 ;
With derived Table
Select a.field1, a.field2, b.field3 From ( Select field1, field2, sum(a.field3) as field3 From Table1 Where field1=123 And field2=345 Group by 1,2 )a Left outer join Table2 b On a.field1=b.field1 And b.field2=b.field2 ;
This technique helped me most of the time (I am dealing a lot with Front end tools) in combination of he statistics there is even no need to set secondary indices.
There are many more to mention, but I am sure you will receive additional comments from other forum members :)
|Copyright 2016 - All Rights Reserved|
|Last Modified: 27 Dec 2016|