On 14 Sep 2011, at 20:45, Brian Fehrle wrote:
>> That is only about 1/30th of your table. I don't think a seqscan makes sense here unless your data is distributed
badly.
>>
> Yeah the more I look at it, the more I think it's postgres _thinking_ that it's faster to do a seqential scan. I'll
beplaying with the random_page_cost that Ondrej suggested, and schedule a time where I can do some explain analyzes
(productionserver and all).
Before you do that, turn off seqscans (there's a session option for that) and see if index scans are actually faster.
Alban Hertroys
--
If you can't see the forest for the trees,
cut the trees and you'll see there is no forest.