Message-ID: <9d4c7b0de250db9aef716600ca6e39652738ff4e.camel@cybertec.at>
Subject: Re: IS NOT DISTINCT FROM statement
From: Laurenz Albe <laurenz.albe@cybertec.at>
To: Artur =?UTF-8?Q?Zaj=C4=85c?= <azajac@ang.com.pl>,
	pgsql-performance@lists.postgresql.org
Date: Fri, 08 Mar 2019 12:45:39 +0100
In-Reply-To: <012a01d4d5a2$5cde0760$169a1620$@ang.com.pl>
References: <012a01d4d5a2$5cde0760$169a1620$@ang.com.pl>
Organization: Cybertec
Content-Type: text/plain; charset="UTF-8"
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

Artur Zając wrote:
> Is there any reason that „NOT DISTINCT FROM” can’t be autotransformed to „=” when value
> on right side of expression is not NULL or is this any way to use index with „IS NOT DISTINCT FROM” statement?

That would subtly change the semantics of the expression:

test=> SELECT NULL IS NOT DISTINCT FROM 21580;
 ?column? 
----------
 f
(1 row)

test=> SELECT NULL = 21580;
 ?column? 
----------
 
(1 row)

One expression is FALSE, the other NULL.

It doesn't matter in the context of your specific query, but it could matter.

Yours,
Laurenz Albe
-- 
Cybertec | https://www.cybertec-postgresql.com