From: Tom Lane <tgl@sss.pgh.pa.us>
To: 1482694023@qq.com
cc: pgsql-bugs@lists.postgresql.org
Subject: Re: BUG #19722: Window PARTITION BY numeric treats equal values with
 different scales as separate partitions
In-reply-to: <19722-1b775eae8675c158@postgresql.org>
References: <19722-1b775eae8675c158@postgresql.org>
Comments: In-reply-to PG Bug reporting form <noreply@postgresql.org>
	message dated "Sat, 26 Sep 2026 08:55:20 -0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2160450.1790433575.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Sat, 26 Sep 2026 10:39:35 -0400
Message-ID: <2160451.1790433575@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2160451.1790433575%40sss.pgh.pa.us>
Precedence: bulk

PG Bug reporting form <noreply@postgresql.org> writes:
> CREATE TEMP TABLE decimal_source (raw_value text NOT NULL);
> INSERT INTO decimal_source VALUES
>     ('1.0'), ('1.00'), ('1.000');

> SELECT amount, partition_size
> FROM (
>     SELECT raw_value::numeric AS amount,
>            COUNT(*) OVER (PARTITION BY raw_value::numeric) AS partition_=
size
>     FROM decimal_source
> ) AS windowed
> WHERE scale(amount) =3D 1;

This is a near-duplicate of many recent reports[1][2][3][4][5].
Not one of them has offered a reason why they think that testing
scale() of a grouped numeric column is a useful or even well-defined
thing to do.  Do you have a real-world use case?  What is it?

			regards, tom lane

[1] https://www.postgresql.org/message-id/flat/CA%2BCOZaA0tnyza0pHx6-UEdM5=
dcLGP4RJJ0KOHFq%2BwVQCY2Jyiw%40mail.gmail.com
[2] https://www.postgresql.org/message-id/flat/19588-e32b82433b3660ea%40po=
stgresql.org
[3] https://www.postgresql.org/message-id/flat/19619-fc646db1c6b2dc90%40po=
stgresql.org
[4] https://www.postgresql.org/message-id/flat/19697-6e7ccba388bbe858%40po=
stgresql.org
[5] https://www.postgresql.org/message-id/flat/19713-635b8656d07464df%40po=
stgresql.org