Archives of the TeradataForum
Message Posted: Tue, 02 Jul 2002 @ 15:26:53 GMT
| Subj: | | Re: V2R3 - V2R4 Migration. |
| |
| From: | | David Picek |
We just moved our V2R4 into production over this past weekend. During the testing we did find problems with our BO queries
running out of spool space. I could never set the spool space for the user high enough to let this query run. The explains show a
totally different path for getting the data which uses up all of the spool space. An interesting way I got around one of the queries
was to actually add a table and a join to the query. This caused the optimizer to take a different path in loading the data into the
spool which resulted in the query working on V2R4. We have a level 2 incident in at NCR regarding this issue. Our rep said that they
recognize this as a problem and are working on a fix. Not sure how long before it gets fixed.
Our rep also mentioned that it is due to the statistics (see thread on "Anyone noticing increased spool usage under V2R412?").
From what I understand, one issue is that if you join a table with a small number of rows with that has a large number of rows the
optimizer doesn't realize that it's going to pull in the large number of rows into the spool. So it takes an inappropriate path and
tries to load the large table into the spool.
I am currently looking at a couple of more queries that have been brought to my attention since our move to production.
David Picek
Data Analyst, Zurich Life US
| 
