OK, two things. First the row estimate starts going way off around the time it gets to the hash aggregate / nested loop which seems to be making the planner use a bad plan for this many rows. You can try issuing
set enable_nestloop = off; before running the query and see if that makes it any faster. Secondly, the first time you run this query you are reading the 1.8G table sequentially, and at about 55MB/s, which isn't gonna get faster without more / faster drives under your machine. On Thu, Aug 6, 2009 at 12:50 AM, Ip Wing Kin John<[email protected]> wrote: > Here u go. Both in the same file. > > On Thu, Aug 6, 2009 at 4:48 PM, Scott Marlowe<[email protected]> wrote: >> Much better... Looks like I got the second one... >> >> Can I get the first one too? Thx. >> >> On Thu, Aug 6, 2009 at 12:46 AM, Ip Wing Kin John<[email protected]> wrote: >>> Hope you can get it this time. >>> >>> John >>> >>> On Thu, Aug 6, 2009 at 4:34 PM, Scott Marlowe<[email protected]> >>> wrote: >>>> Sorry man, it's not coming through. Try it this time addressed just to me. >>>> >>>> On Thu, Aug 6, 2009 at 12:23 AM, Ip Wing Kin John<[email protected]> >>>> wrote: >>>>> Hi scott >>>>> >>>>> I attached the query plan with this email. The top one is the first >>>>> run after I restarted my machine. And the bottom one is the second >>>>> run. >>>>> >>>>> I am using PostgreSQL 8.3 on Solaris 10. >>>>> >>>>> cheers >>>>> >>>>> On Thu, Aug 6, 2009 at 4:15 PM, Scott Marlowe<[email protected]> >>>>> wrote: >>>>>> On Wed, Aug 5, 2009 at 11:21 PM, <[email protected]> wrote: >>>>>>> Sorry post again. >>>>>> >>>>>> Nope, still mangled. Can you attach it? >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> John >>>>> >>>> >>>> >>>> >>>> -- >>>> When fascism comes to America, it will be intolerance sold as diversity. >>>> >>> >>> >>> >>> -- >>> John >>> >> >> >> >> -- >> When fascism comes to America, it will be intolerance sold as diversity. >> > > > > -- > John > -- When fascism comes to America, it will be intolerance sold as diversity. -- Sent via pgsql-performance mailing list ([email protected]) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-performance
