Archives of the TeradataForum
Message Posted: Wed, 09 Mar 2005 @ 17:06:19 GMT
| Subj: || || Re: Multi Column Statistics Puzzle |
| From: || || Victor Sokovin |
| ||Does anyone know why this may happen? I have a simple query below. I ran an explain on the query and then ran the query. I was then
trying to improve the speed, so I collected statistics on both columns in the WHERE clause. I then ran the explain again, but instead of being
"high confidence" as in the first explain (pre STATS), it came back "low confidence". The number of rows that were returned was 3 (as in the
second explain plan), but I am confused as to the confidence level.|| |
I guess everybody would be.
I'm sure I've seen these tables before but can't recall their definitions yet. Almost an obligatory question: do you happen to have SI on one
or both of the columns that you use in the query? I've been told on a few occasions that SI should not matter at all in this context but I am
still not convinced and think of them as first suspects. If you don't have SI, then I'll have to pass on this one. Sorry.