Are you using any trunc functions while querying.
Do this. SELECT column_name,count(*) from small_table group by column_name;
Run this for all the tables. My guess is that atleast 30% of the rows in the
small table is for the value that you gave. So the optimizer is forced to read
the index
I have five tables, three are 3 million rows, one is a little more than 1
million, one a little less than one million.
All tables are structured the same, with the same indexing, just different
types of data. There is a date column
with a normal index.
A query with a 'between' on the date
- Original Message -
Why would taking an index off make this query faster? What doesn't the CBO
know that not using an index
is the best path?
The physical sequence of the rows in the smaller table
vis-a-vis the range you're querying on.
Cheers
Nuno Souto
[EMAIL PROTECTED]
--