On Wed, Apr 27, 2011 at 9:22 AM, Claudio Freire <klaussfreire@gmail.com> wrote: > The problem there, I think, is that the planner is doing a full join, > instead of a semi-join. Or, rather, computing cost as if it was a full join. I'm not sure why.
pgsql-performance by date:
Соглашаюсь с условиями обработки персональных данных