Re: [HACKERS] current- crash - Mailing list pgsql-hackers

From Bruce Momjian
Subject Re: [HACKERS] current- crash
Date
Msg-id 199809222201.SAA05135@candle.pha.pa.us
Whole thread Raw
In response to Re: [HACKERS] current- crash  (Michael Reifenberger <root@nihil.plaut.de>)
Responses datetime questions and fix to dt.c
List pgsql-hackers
> On Tue, 22 Sep 1998, Bruce Momjian wrote:
> ...
> > > I tried your script but I can't see nothing wrong. :)
> Right!
> The crash is fixed.
> Thanks Bruce!
> Now, shouldn't I see Index Scans below or is it a optimizer feature
> because the tables are too empty?
> > > Merge Join  (cost=0.00 size=1 width=8)
> > >   ->  Seq Scan  (cost=0.00 size=0 width=0)
> > >         ->  Sort  (cost=0.00 size=0 width=0)
> > >               ->  Seq Scan on a  (cost=0.00 size=0 width=4)
> > >   ->  Seq Scan  (cost=0.00 size=0 width=0)
> > >         ->  Sort  (cost=0.00 size=0 width=0)
> > >               ->  Seq Scan on b  (cost=0.00 size=0 width=4)

Not sure.  Joins usually don't use index scans if the table is used in
the outer part of the join loop because it is just quicker to sort them.
Also the tables must have a significant amount of data in them to use
indexes.

I am also not sure how the system handles an OR and a join on the same
column.  I would think it does not use an index, but perhaps it does the
restriction using an index, then does the join.  That is what it should
do.

--
Bruce Momjian                          |  830 Blythe Avenue
maillist@candle.pha.pa.us              |  Drexel Hill, Pennsylvania 19026
http://www.op.net/~candle              |  (610) 353-9879(w)
  +  If your life is a hard drive,     |  (610) 853-3000(h)
  +  Christ can be your backup.        |

pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: NOT boolfield kills backend
Next
From: "Thomas G. Lockhart"
Date:
Subject: Re: NOT boolfield kills backend