Archives of the TeradataForum
Message Posted: Tue, 28 Oct 2003 @ 17:01:33 GMT
Has anyone done any testing to ascertain whether it's best to leave secondary indexes in place when inserting a significant number of rows into an existing table, or whether it's better to Drop and Re-create the index ?
Prior to V2R5 we were certainly of the opinion that the latter approach was better.
Unfortunately my testing appears very inconclusive and maybe suffering from varying loads on the box whilst my testing proceeds.
I'd be using an INSERT/SELECT approach with around 20,000 rows being added to a table of tens of millions.
Any clear conclusions that anyone's been able to derive would be useful.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 23 Jun 2019|