Monetary Data Types Improvement
Hackorum builds and tests every patch posted to the lists, not only commitfest submissions. This is Hackorum's own CI rather than the PostgreSQL project's, and it is still under testing - please report anything that looks wrong.
You can run a PostgreSQL built from this patch straight from Docker, with no checkout and no build:
docker run --rm -p 5432:5432 ghcr.io/hackorum-dev/postgres-patch:t77597psql -h localhost -U postgresBuilt from patchset v3 (message #3), September 21, 2026 at 04:41 AM.
Every patchset is also pushed to a branch of our PostgreSQL fork, so you can check out the same tree CI built. Without a PostgreSQL checkout:
git clone --branch t77597_3 https://github.com/hackorum-dev/postgres.gitIn a checkout you already have, add the fork once:
git remote add hackorum https://github.com/hackorum-dev/postgres.gitthen, for this patchset and every later one:
git fetch hackorum t77597_3 && git checkout t77597_3Patchset v3 (message #3) is on t77597_3
The following documentation comment has been logged on the website:
Page: https://www.postgresql.org/docs/14/datatype-money.html
Description:
It's not explicitly obvious that money doesn't behave like a normal numeric
type in that executing a procedure with a negative numeric value for money
causes an error. The solution to this is to pass the value as a string. For
example, -15.99 for money should be expressed as '-15.99'.
On 2024-03-17 03:16 +0100, PG Doc comments form wrote:
The following documentation comment has been logged on the website:
Page: https://www.postgresql.org/docs/14/datatype-money.html
Description:It's not explicitly obvious that money doesn't behave like a normal numeric
type in that executing a procedure with a negative numeric value for money
causes an error. The solution to this is to pass the value as a string. For
example, -15.99 for money should be expressed as '-15.99'.
I assume it's intended to be used like other numeric datatypes (although
with known issues[1]https://wiki.postgresql.org/wiki/Don%27t_Do_This#Don.27t_use_money) given that this page does not mention any
operations besides division.
I see no reason why unary minus and even unary plus shouldn't be
implemented if negative amounts are already possible. Maybe it's not
worth the effort if one can just do (OP x::numeric)::money for any
unary numeric operator OP instead. CREATE OPERATOR is another option.
Maybe add a note like:
"Money does not implement all operators that one might expect of a
numeric type. For example, use (-amount::money::numeric)::money to
negate amount."
That would also fit nicely with the existing examples on casting to
numeric and float8. The attached patch does that.
[1]: https://wiki.postgresql.org/wiki/Don%27t_Do_This#Don.27t_use_money
--
Erik
On Tue, 19 Mar 2024 at 07:43, Erik Wienhold <ewie@ewie.name> wrote:
Maybe add a note like:
"Money does not implement all operators that one might expect of a
numeric type. For example, use (-amount::money::numeric)::money to
negate amount."That would also fit nicely with the existing examples on casting to
numeric and float8. The attached patch does that.
My vote would go to adding a deprecation notice to that section of the
docs. There's some talk [1]/messages/by-id/ZXgh74Ykj3iWvXKr@paquier.xyz about how we discourage the usage of the
money type and that goes on to discuss the possibilities of moving it
into a contrib module.
My hope would be that deprecation notice would steer most people away
from using it and therefore reduce the number of questions about it
due to fewer new use cases of it.
David
On 2024-03-18 23:24 +0100, David Rowley wrote:
My vote would go to adding a deprecation notice to that section of the
docs. There's some talk [1] about how we discourage the usage of the
money type and that goes on to discuss the possibilities of moving it
into a contrib module.My hope would be that deprecation notice would steer most people away
from using it and therefore reduce the number of questions about it
due to fewer new use cases of it.
+1
Huh, I didn't know that it used to have a deprecation notice at some
point. But that note was removed in 8.3:
/messages/by-id/20071009123315.5fb283c1.darcy@druid.net
--
Erik
On Wed, 20 Mar 2024 at 10:59, Erik Wienhold <ewie@ewie.name> wrote:
On 2024-03-18 23:24 +0100, David Rowley wrote:
My vote would go to adding a deprecation notice to that section of the
docs. There's some talk [1] about how we discourage the usage of the
money type and that goes on to discuss the possibilities of moving it
into a contrib module.My hope would be that deprecation notice would steer most people away
from using it and therefore reduce the number of questions about it
due to fewer new use cases of it.+1
Huh, I didn't know that it used to have a deprecation notice at some
point. But that note was removed in 8.3:
Sadly that was a mistake. Money is not really a useful type.
Dave
On 3/20/24 11:07 AM, Dave Cramer wrote:
On Wed, 20 Mar 2024 at 10:59, Erik Wienhold <ewie@ewie.name
<mailto:ewie@ewie.name>> wrote:On 2024-03-18 23:24 +0100, David Rowley wrote:
My vote would go to adding a deprecation notice to that section
of the
docs. There's some talk [1] about how we discourage the usage of the
money type and that goes on to discuss the possibilities of moving it
into a contrib module.My hope would be that deprecation notice would steer most people away
from using it and therefore reduce the number of questions about it
due to fewer new use cases of it.[1]
/messages/by-id/ZXgh74Ykj3iWvXKr@paquier.xyz
</messages/by-id/ZXgh74Ykj3iWvXKr@paquier.xyz>+1
Huh, I didn't know that it used to have a deprecation notice at some
point. But that note was removed in 8.3:/messages/by-id/20071009123315.5fb283c1.darcy@druid.net </messages/by-id/20071009123315.5fb283c1.darcy@druid.net>
Sadly that was a mistake. Money is not really a useful type.
Here's[1]/messages/by-id/18240-c5da758d7dc1ecf0@postgresql.org the latest "let's remove money" discussion (there's allegedly
a hackers thread too, but I'm having trouble finding it.
+1 on reinstating the deprecation notice, given it'll be some years
before we can fully remove it (at least based on the discussion).
Thanks,
Jonathan
[1]: /messages/by-id/18240-c5da758d7dc1ecf0@postgresql.org
/messages/by-id/18240-c5da758d7dc1ecf0@postgresql.org
Coming from using SQL Server, money was the easiest way I saw to replicate FORMAT(<value>,'C2') for reporting purposes but then I guess I could have just cast it as money for the report/view. Beyond that, using NUMERIC(<num>,2) works fine.
Brian Powell Jr, EI
-----Original Message-----
From: Jonathan S. Katz <jkatz@postgresql.org>
Sent: Thursday, March 21, 2024 3:12 PM
To: Dave Cramer <davecramer@postgres.rocks>; Erik Wienhold <ewie@ewie.name>
Cc: David Rowley <dgrowleyml@gmail.com>; b.powell.jr@outlook.com; pgsql-docs@lists.postgresql.org
Subject: Re: Monetary Data Types Improvement
On 3/20/24 11:07 AM, Dave Cramer wrote:
On Wed, 20 Mar 2024 at 10:59, Erik Wienhold <ewie@ewie.name
<mailto:ewie@ewie.name>> wrote:On 2024-03-18 23:24 +0100, David Rowley wrote:
My vote would go to adding a deprecation notice to that section
of the
docs. There's some talk [1] about how we discourage the usage of the
money type and that goes on to discuss the possibilities of moving it
into a contrib module.My hope would be that deprecation notice would steer most people away
from using it and therefore reduce the number of questions about it
due to fewer new use cases of it.[1]
/messages/by-id/ZXgh74Ykj3iWvXKr@paquier.xyz
</messages/by-id/ZXgh74Ykj3iWvXKr@paquier.xyz>
+1
Huh, I didn't know that it used to have a deprecation notice at some
point. But that note was removed in 8.3:/messages/by-id/20071009123315.5fb283c1.darcy@
druid.net
</messages/by-id/20071009123315.5fb283c1.darcy%254
0druid.net>Sadly that was a mistake. Money is not really a useful type.
Here's[1]/messages/by-id/18240-c5da758d7dc1ecf0@postgresql.org the latest "let's remove money" discussion (there's allegedly a hackers thread too, but I'm having trouble finding it.
+1 on reinstating the deprecation notice, given it'll be some years
before we can fully remove it (at least based on the discussion).
Thanks,
Jonathan
[1]: /messages/by-id/18240-c5da758d7dc1ecf0@postgresql.org
/messages/by-id/18240-c5da758d7dc1ecf0@postgresql.org