Re: [BUGS] BUG #2930: Hash join abyssmal with many null fields.

Tom Lane wrote: Maciej Babinski <[EMAIL PROTECTED]> writes: Tom Lane wrote: I see no bug here. AFAICT your "much faster" query gets that way by having eliminated all the candidate join rows on the B side. The additional clause eliminates no rows beyond what the existing clause would. Any r

Re: [BUGS] BUG #2930: Hash join abyssmal with many null fields.

Tom Lane wrote: "Maciej Babinski" <[EMAIL PROTECTED]> writes: Hash join of columns with many null fields is very slow unless the null fields are commented out. I see no bug here. AFAICT your "much faster" query gets that way by having eliminated all the candidate join rows on the B side.

Re: [BUGS] BUG #2930: Hash join abyssmal with many null fields.

Maciej Babinski <[EMAIL PROTECTED]> writes: > Tom Lane wrote: >> I see no bug here. AFAICT your "much faster" query gets that way by >> having eliminated all the candidate join rows on the B side. > The additional clause eliminates no rows beyond what the existing > clause would. Any row elimina

Re: [BUGS] BUG #2930: Hash join abyssmal with many null fields.

"Maciej Babinski" <[EMAIL PROTECTED]> writes: > Hash join of columns with many null fields is very slow unless the null > fields are commented out. I see no bug here. AFAICT your "much faster" query gets that way by having eliminated all the candidate join rows on the B side.