Archives of the TeradataForum
Message Posted: Fri, 11 Mar 2005 @ 18:17:36 GMT
| Subj: || || Re: Restrict number of Rows Affected |
| From: || || Stover, Terry |
If you're deleting a substantial percentage of the data you can also do an insert into a new table using the appropriate filter, then rename
the tables. The insert into the empty table isn't logged and you don't get the dupe PI checking. If you have the space this is the quickest way
to do it. Multiload deletes are pretty good, we haven't done many, but it was easily 5-10 times faster than a simple sql delete.
Looping through a series of delete statements isn't going to save any time, it only helps if you're running out of space for the transient
journal. I've done a similar thing with sql server where we looped over one column of the primary key.