Heikki Linnakangas <[EMAIL PROTECTED]> writes:
> Tom Lane wrote:
>> Comments, better ideas?
> How about building a separate Var-node for the variable when it's above
> an outer join?
[ itch... ] That would be a whole lot *more* invasive than what I'm
proposing now. It might be an interesting a
Tom Lane wrote:
Comments, better ideas?
How about building a separate Var-node for the variable when it's above
an outer join? That node would point to the original Var, and have an
additional modifier which indicates the percentage of injected nulls.
--
Heikki Linnakangas
EnterpriseDB
I've been thinking about what it will take to solve the problem noted here:
http://archives.postgresql.org/pgsql-performance/2007-05/msg00325.php
which briefly is that 8.2 is really bad at estimating the number of
rows returned by locutions like
SELECT ... FROM
tab1 LEFT JOIN tab
