|
|
Archives of the TeradataForum
Message Posted: Fri, 14 Nov 2003 @ 20:36:17 GMT
Subj: | | Re: Trace in Teradata |
|
From: | | Ballinger, Carrie |
Milan,
If you are on V2R5 you can enable query logging for the DBQLStepTbl, which can help with query tuning. When you log to the step table,
you will get one row for each step that executed in your query, with the start and end times and row counts for each step. Here's what
some sample output for one query looks like. You can see in this case that almost all the query time was spent in step 4, the join, and
that 131,158 rows were spooled in that step.
QueryID Step StepName StepStartDate StepStopDate RowCount
765 1 MLK 2003-06-27 08:38:02 2003-06-27 08:38:02 1
765 2 MLK 2003-06-27 08:38:02 2003-06-27 08:38:02 1
765 3 MLK 2003-06-27 08:38:02 2003-06-27 08:38:02 1
765 4 JIN 2003-06-27 08:38:02 2003-06-27 08:47:11 131158
765 5 SUM 2003-06-27 08:47:11 2003-06-27 08:47:13 1
765 6 RET 2003-06-27 08:47:13 2003-06-27 08:47:13 5
765 7 Edt 2003-06-27 08:47:13 2003-06-27 08:47:13 1
765 8 RSF ? ? ?
If you are using query log for query debugging, you might want to end logging as soon as the query completes, to get the cache to write
its contents to disk so you can see the rows. Otherwise you'll have to wait until the cache fills up and it forces the rows out, before you
can see them.
Thanks, --Carrie
| |