Many thanks! Ismo's reply solved the problem, and Tom's reply solved
another one.
Marko
Ismo Tuononen wrote:
select studentid,max(score) from studentprofile group by studentid;
On 20/02/07, Tom Lane <[EMAIL PROTECTED]> wrote:
The optimal index for this would be on studentprofile(studentid,s
Did you run ANALYZE on your data after importing it into 8.2.3? Is there an
index on the seqfil column? If so, you should post the output of EXPLAIN
ANALYZE from both systems if possible.
> -Original Message-
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On Behalf Of
> Glenn S
On Tue, 20 Feb 2007, Glenn Sullivan wrote:
I am updating from 7.4.5 to 8.2.3. I have noticed a significant
slowdown in simple searches such as
"select filename from vnmr_data where seqfil = 'sems';"
This returns 12 rows out of 1 million items in the table.
On 7.4.5, this takes about 1.5 seco
I am updating from 7.4.5 to 8.2.3. I have noticed a significant
slowdown in simple searches such as
"select filename from vnmr_data where seqfil = 'sems';"
This returns 12 rows out of 1 million items in the table.
On 7.4.5, this takes about 1.5 seconds. On 8.2.3, it is taking
about 9 seconds
It's not necessarily the join order that's an issue; it could also be
due to the merge join that it does in the first case. I've also run into
situations where the cost estimate for a merge join is way off the mark.
Rather than forcing the join order, you might try setting
enable_mergejoin=false.
