On Wed, 1 May 2002 postgres@vrane.com wrote:
> On Wed, May 01, 2002 at 11:54:37PM +0100, Nigel J. Andrews wrote:
> >
> > On Wed, 1 May 2002 postgres@vrane.com wrote:
> >
> > Then run
> > EXPLAIN ANALYZE SELECT * FROM TABLE <mytable> WHERE <booleanfield> = true;
> >
> > Then run
> > UPDATE <mytable> SET <booleanfield> = false WHERE <booleanfield> = false;
> >
> > (several times)
> >
> > What you should see is a seqscan in both plans with the final number of rows
> > the same but the second will have taken longer.
>
> I haven't tried it yet but isn't creating an index here
> the logical thing to do here. Whenever I have "where"
> clause in my queries I always create an index.
Damn, knew I'd miss something out. Yes, it would seem a good idea to create an
index on that boolean column. I'd still expect the second query (in the explain
analyze) to take longer to complete. I think I might give this a go myself
sometime just to check.
--
Nigel J. Andrews
Director
---
Logictree Systems Limited
Computer Consultants