From: Tom Lane <tgl@sss.pgh.pa.us>
To: Rajkumar Raghuwanshi <rajkumar.raghuwanshi@enterprisedb.com>
cc: pgsql-hackers <pgsql-hackers@postgresql.org>
Subject: Re: negative bitmapset member not allowed Error with partition
 pruning
In-reply-to: 
 <CAKcux6k=udMa_oOXnbHduBwvJ8SnUaQxRLPA=D2+ryYoyT6eeA@mail.gmail.com>
References: 
 <CAKcux6k=udMa_oOXnbHduBwvJ8SnUaQxRLPA=D2+ryYoyT6eeA@mail.gmail.com>
Comments: In-reply-to Rajkumar Raghuwanshi
 <rajkumar.raghuwanshi@enterprisedb.com>
	message dated "Thu, 26 Jul 2018 15:40:01 +0530"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <16589.1532622503.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Thu, 26 Jul 2018 12:28:23 -0400
Message-ID: <16590.1532622503@sss.pgh.pa.us>
Precedence: bulk

Rajkumar Raghuwanshi <rajkumar.raghuwanshi@enterprisedb.com> writes:
> I am getting "ERROR:  negative bitmapset member not allowed" when
> enable_partition_pruning set to true with below test case.

Confirmed here.  It's failing in perform_pruning_combine_step,
which reaches this:

        result->bound_offsets =3D bms_add_range(NULL, 0, boundinfo->ndatum=
s - 1);

with boundinfo->ndatums =3D=3D 0.  It's not clear to me whether that situa=
tion
should be impossible or not.  If it is valid, perhaps all we need is
something like

        if (boundinfo->ndatums > 0)
            result->bound_offsets =3D bms_add_range(NULL, 0, boundinfo->nd=
atums - 1);
        else
            result->bound_offsets =3D NULL;

although that then opens the question of whether downstream code is
OK with bound_offsets being empty.

(BTW, I'm not sure that it was wise to design bms_add_range to fail for
empty ranges.  Maybe it'd be better to redefine it as a no-op for
upper < lower?)

			regards, tom lane