Message-ID: <c8c5090a0d720db8a5c4c382cce59afde3dc65d2.camel@cybertec.at>
Subject: Re: BUG #18976: -0.0 with float8 will be transformed to 0 in
 prepare statement but not in normal execution
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: 798604270@qq.com, pgsql-bugs@lists.postgresql.org
Date: Thu, 03 Jul 2025 14:23:42 +0200
In-Reply-To: <18976-c1bc92542b90e5e6@postgresql.org>
References: <18976-c1bc92542b90e5e6@postgresql.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.56.2 (3.56.2-1.fc42) 
MIME-Version: 1.0
Archived-At: 
 <https://www.postgresql.org/message-id/c8c5090a0d720db8a5c4c382cce59afde3dc65d2.camel%40cybertec.at>
Precedence: bulk

On Thu, 2025-07-03 at 03:03 +0000, PG Bug reporting form wrote:
> The following bug has been logged on the website:
>=20
> Bug reference:      18976
> Logged by:          Chi Zhang
> Email address:      798604270@qq.com
> PostgreSQL version: 18beta1
> Operating system:   Ubuntu 24.04 and docker
> Description:       =20
>=20
> The value -0.0 with float8 is transformed to 0 in prepared statements but
> remains -0 in normal execution. Although 0 and -0 are numerically equal,
> this discrepancy can lead to subtle bugs in certain cases=E2=80=94for exa=
mple, when
> the value is cast to a VARCHAR, as illustrated below.
>=20
> PREPARE prepare_query (float8) AS SELECT CAST($1 AS VARCHAR) =3D
> CAST(-0.0::float8 AS VARCHAR);
> EXECUTE prepare_query(-0.0); -- f

That's not a bug, but a pilot error.  If you feed a "float8", the result is=
t TRUE:

EXECUTE prepare_query(-0.0::float8);

 ?column?=20
=E2=95=90=E2=95=90=E2=95=90=E2=95=90=E2=95=90=E2=95=90=E2=95=90=E2=95=90=E2=
=95=90=E2=95=90
 t
(1 row)

Yours,
Laurenz Albe