To: "Guillaume Smet" <guillaume.smet@gmail.com>
cc: pgsql-performance@postgresql.org
Subject: Re: Question about Bitmap Heap Scan/BitmapAnd 
In-reply-to: <1d4e0c10702130832i35484abfmb229f5d0b4d9223b@mail.gmail.com> 
References: <1d4e0c10702130832i35484abfmb229f5d0b4d9223b@mail.gmail.com>
Comments: In-reply-to "Guillaume Smet" <guillaume.smet@gmail.com>
	message dated "Tue, 13 Feb 2007 17:32:58 +0100"
Date: Tue, 13 Feb 2007 11:49:13 -0500
Message-ID: <1285.1171385353@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

"Guillaume Smet" <guillaume.smet@gmail.com> writes:
> What surprises me is that "parking" is in the filter and not in the
> Recheck Cond whereas it's part of the second Bitmap Index Scan of the
> Bitmap And node.

That's probably because of this:

    /*
     * When dealing with special or lossy operators, we will at this point
     * have duplicate clauses in qpqual and bitmapqualorig.  We may as well
     * drop 'em from bitmapqualorig, since there's no point in making the
     * tests twice.
     */
    bitmapqualorig = list_difference_ptr(bitmapqualorig, qpqual);

What's not immediately clear is why the condition was in both lists to
start with.  Perhaps idx_lieu_parking is a partial index with this as
its WHERE condition?

			regards, tom lane