Archives of the TeradataForum
Message Posted: Thu, 02 Dec 1999 @ 08:03:43 GMT
One technique that has been used (and this in not the end-all-be-all) but works quite well would be to create a temp table (i.e. CREATE TABLE with a different name but same EXACT structure - i.e. same Primary Index - but with no secondary indexes) and INSERT INTO that table as SELECT * and conditions that will not pull over the rows you want to 'delete'. This operation is performed in parallel and all the rows stay on the same VAMPs (quite fast). After the INSERT is completed, create the 7 indexes in parallel then drop the 'old table' and rename the new table as the old name. Set up a test and see if this works a bit faster. If you are doing this during 'user time', this facilitates user access for a longer time. also, if you are using Views (i.e. 1:1) to access the base tables, you might want to also use LOCKING FOR ACCESS in your view definitions.
Doing Updates/Deletes of that many rows on a table that has 7 indexes is quite costly. You have to delete all the index entries whilst deleting the row itself. You can email me direct to discuss - get examples, etc. if you would like.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|