On Mon, 5 Dec 2005, Tom Lane wrote:
>
> Hm, I can't duplicate this in a simple test...
> Can you try updating to 8.1 branch tip ...
> Actually, a quick and dirty thing would be to try my would-be test case
> below, and see if you get a seqscan on your copy.
With your simple test-case I did not ge
I have a case where an outer join's taking 10X more time than
a non-outer join; and it looks to me like the outer join could
have taken advantage of the same indexes that the non-outer join did.
In both cases, the outermost thing is a nested loop. The
top subplan gets all "point features" whre f
Can anyone please help me make my JOIN find the right index to use?
It seems strange to me that in the two queries listed below, the
LEFT OUTER JOIN can find the most efficient index to use, while
the unadorned JOIN can not. The result is that my query is
orders of magnitude slower than it s
