On Fri, May 18, 2007 at 02:51:42PM +0300, Liviu Ionescu wrote:
> it helped, the new version of the query takes 2303 ms on both 8.1.4 and 8.2.4.
And the old one?
> any idea why the 8.2.4 planner is not happy with the initial select? was it
> just a big chance that it worked in 8.1.4 or the 8.2.4 planner has a
> problem?
I guess it was more or less by chance, especially as 8.1 did not reorder
outer joins. Others might know more about the estimation, though.
/* Steinar */
--
Homepage: http://www.sesse.net/
