Archives of the TeradataForum
Message Posted: Wed, 23 Mar 2005 @ 13:11:04 GMT
The fastest way to do this (if you have the space to support this effort) is as follows:
Backing up the data before purge:
1. Create an Empty Table for the rows you wish to archive using a naming standard that defines the dated data in the table (ie: Jan2005_DateID).
2. Locking table databasename.tablename for Access Insert/Select the defined rows into the empty table where DATE_ID = '200501'.
3. Archive the table
4. DROP TABLE once archive is complete
Removing the Rows from the Table: (Here is where you need to determine if a straight delete or an insert/select is faster):
1. Create table databasename.tablename_TMP as databasename.tablename with no data;
2. Insert into databasename.tablename_TMP Select from databasename.tablename where NOT DATE_ID = '200501';
3. RENAME TABLE databasename.tablename to databasename.tablename_OLD;
4. RENAME TABLE databasename.tablename_TMP to databasename.tablename;
5. COLLECT STATS ON databasename.tablename (define stats);
See if this helps your effort.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|