On Tue, 26 Apr 2005, Tom Lane wrote:
> Pavel Stehule <[EMAIL PROTECTED]> writes:
> > I tested bitmap scan and maybe I didnt find good examples, but with bitmap
> > scan is slower than hashjoin. Only when I use non otiptimized SELECT bps
> > was little bit faster. All my SELECTs are equal.
>
> B
It's interesting, that Tom's example behaves different on my notebook:
8.02 (default optimization)
regression=# explain analyze select * from tenk1 where hundred between 1 and 10 and thousand between 1 and 100;
QUERY PLAN
I didn't vacuum tenk1 after loading into 8.02, so optimizer was optimistic
and used index. After vacuuming I got what's Tom get.
On Tue, 26 Apr 2005, Tom Lane wrote:
In 8.0 this looks like
regression=# explain analyze select * from tenk1 where hundred between 1 and 10
and thousand between 1 and 10
Pavel Stehule <[EMAIL PROTECTED]> writes:
> I tested bitmap scan and maybe I didnt find good examples, but with bitmap
> scan is slower than hashjoin. Only when I use non otiptimized SELECT bps
> was little bit faster. All my SELECTs are equal.
Bitmap scans can't possibly be any faster for cases
Hello,
I tested bitmap scan and maybe I didnt find good examples, but with bitmap
scan is slower than hashjoin. Only when I use non otiptimized SELECT bps
was little bit faster. All my SELECTs are equal.
bsp off bsp on (ms)
SELECT 1 39.798 37.993
SELECT 2 0.310
