On 21/07/2012 20:19, Claudio Freire wrote:
On Sat, Jul 21, 2012 at 4:16 PM, Ioannis Anagnostopoulos
<ioannis@xxxxxxxxxx> wrote:
I am not sure that I can see an improvement, at least on src_id that have
lots of msg_id per day the query never returned even 5 hours later running
"exaplain analyze". For smaller src_id
(message wise) there might be some improvement or it was just the analyse
that I run. As I said the stats goes quickly out of scope because of the big
number of updates. So it looks like that
it is not the "funny" "where" concatenation or some kind of index
construction problem. Which brings us back to the issue of the
"statistics_target" on per column. My problem is that given the
query plan I provided you yesterday, I am not sure which columns
statistics_target to touch and what short of number to introduce. Is there
any rule of thumb?
What's the size of your index, tables, and such?
In GB I mean, not tuples.
The message_copies_wk2 that I currently hit is 13GB and 11 the Indexes, the
ship_a_pos_messages_wk2 is 17GB and 2.5MB the index and the ship_objects
is 150MB table and index approx.
Yiannis
--
Sent via pgsql-performance mailing list (pgsql-performance@xxxxxxxxxxxxxx)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-performance
