Re: Optomizing left outer joins

Josh Berkus <josh(at)agliodbs(dot)com> writes:
> FWIW, outer joins are slower than regular joins on all RDBMSs I've tested --
> often up to 5 times slower. I'm not sure whether this is just the planner
> restricitons inherent in an outer join, or whether this is something about
> the required join algorithm itself.

There isn't anything fundamentally slower about left join vs plain join
(unless, perhaps, the left join has to generate a really large number
of null-extended rows that the plain join would not generate). My money
is on the left join having prevented some optimization that could be
applied in the plain-join case. But since we have not been shown any
exact details of the tables or the query, it's impossible to do more
than wager...

regards, tom lane