Hi!

On Thu, Sep 8, 2011 at 15:11, Marko Tiikkaja
<marko(dot)tiikkaja(at)2ndquadrant(dot)com> wrote:
> Here's a patch for $SUBJECT based on the feedback here:
> http://archives.postgresql.org/message-id/9053(dot)1295888538(at)sss(dot)pgh(dot)pa(dot)us

I tried this patch and noticed something weird. This is probably not
intentional:

explain analyze select * from generate_series(1,10) i where i>1 and
now()!=now();

Result  (cost=0.01..12.51 rows=333 width=4) (actual time=0.010..0.010
rows=0 loops=1)
  One-Time Filter: (now() <> now())
  ->  Function Scan on generate_series i  (cost=0.01..12.51 rows=333
width=4) (never executed)
        Filter: (i > 1)
        Rows Removed by Filter: -nan

Since it's never executed, nfiltered shows up as "-nan"

Regards,
Marti

In response to

• EXPLAIN and nfiltered, take two at 2011-09-08 12:11:52 from Marko Tiikkaja

Responses

• Re: EXPLAIN and nfiltered, take two at 2011-09-12 09:12:03 from Marko Tiikkaja

pgsql-hackers by date

Next:	From: Jeff Davis	Date: 2011-09-10 16:50:40
	Subject: collation, arrays, and ranges
Previous:	From: Andy Colson	Date: 2011-09-10 16:49:27
	Subject: Re: WARNING: pgstat waiting