BUG #18730: Inequality comparison operators and SMALLINT negative immediate value

Started by PG Bug reporting formover 1 year ago2 messagesbugs
Jump to latest
#1PG Bug reporting form
noreply@postgresql.org

The following bug has been logged on the website:

Bug reference: 18730
Logged by: Nat Makarevitch
Email address: nat@makarevitch.org
PostgreSQL version: 17.2
Operating system: Linux
Description:

Hi!

create table test_smallint(smint smallint);

insert into test_smallint values (-1);

select * from test_smallint where smint<>-2;
smint
═══════
-1
(1 row)

BEWARE: there is a space between '=' and '-':
select * from test_smallint where smint!= -2;
smint
═══════
-1
(1 row)

So far, so good.

BEWARE: there isn't any space between '=' and '-':
select * from test_smallint where smint!=-2;
ERROR: operator does not exist: smallint !=- integer
LINE 1: select * from test_smallint where smint!=-2;
^
HINT: No operator matches the given name and argument types. You might need
to add explicit type casts.

Isn't the operator "!=" theoritically "an alias, which is converted to <> at
a very early stage of parsing" (as per
https://www.postgresql.org/docs/17/functions-comparison.html )?

AFAIK there is no "=-" alias/macro, and I can't see any reason to implicitly
cast "-2" to INTEGER while "2" type stays inferred.

Thank you!

#2David G. Johnston
david.g.johnston@gmail.com
In reply to: PG Bug reporting form (#1)
Re: BUG #18730: Inequality comparison operators and SMALLINT negative immediate value

On Monday, December 2, 2024, PG Bug reporting form <noreply@postgresql.org>
wrote:

The following bug has been logged on the website:

Bug reference: 18730
Logged by: Nat Makarevitch
Email address: nat@makarevitch.org
PostgreSQL version: 17.2
Operating system: Linux
Description:

BEWARE: there isn't any space between '=' and '-':
select * from test_smallint where smint!=-2;
ERROR: operator does not exist: smallint !=- integer
LINE 1: select * from test_smallint where smint!=-2;
^
HINT: No operator matches the given name and argument types. You might
need
to add explicit type casts.

The operator !=- is allowed per [1], since it contains !, so the parser
captures the - as per of the operator as opposed to leaving - as a break
character.

1.
https://www.postgresql.org/docs/current/sql-syntax-lexical.html#SQL-SYNTAX-OPERATORS

Isn't the operator "!=" theoritically "an alias, which is converted to <>
at
a very early stage of parsing" (as per
https://www.postgresql.org/docs/17/functions-comparison.html )?

The != operator and !=- operators are two different operators. Only the
former is converted to <>, and not any symbol subsequence of those two
characters.

David J.