Robert Treat <xzilla@users.sourceforge.net> writes:
> When it happens it tends to look something like this:
> http://archives.postgresql.org/pgsql-performance/2006-01/msg00154.php
> Funny that for all the people who claim that improving the planner should be
> the primary goal that no one ever took interest in the above case.
Well, you didn't provide sufficient data for anyone else to reproduce
the problem ...
regards, tom lane