Archives of the TeradataForum
Message Posted: Sat, 01 Mar 2003 @ 23:16:24 GMT
<-- Anonymously Posted: Saturday, March 01, 2003 17:38 -->
This question is related to the performance between two different methods. I need to create a Dimension table from a Fact table. I will be loading all distinct values of one column into a table. I expect to get maybe 10 rows into the target table from a source table of 100,000 rows. I have different methods and am considering:
1. Do a GROUP BY or Distinct on the source of a INSERT/SELECT. This of course will require the sorting and redistribution and aggregation (ARSA).
2. Make the target table a SET table, do not use any GROUP BY or DISTINCT on the INSERT/SELECT, and let the Duplicate Row check for the SET table weed out the duplicates.
The question is which method would perform the best.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|