Обсуждение: Inconsistent cast to "char"
I noticed that the following does not produce consistent cast results:
select 1::int8::"char", 1::int4::"char", 1::int2::"char"
The following yields 49, 1, 49 respectively:
select 1::int8::"char"::int, 1::int4::"char"::int, 1::int2::"char"::int
regards,
Ian Campbell
"Ian R. Campbell" <ian.campbell@thepathcentral.com> writes: > I noticed that the following does not produce consistent cast results: > select 1::int8::"char", 1::int4::"char", 1::int2::"char" Concretely, that gives: regression=# select 1::int8::"char", 1::int4::"char", 1::int2::"char"; char | char | char ------+------+------ 1 | \x01 | 1 (1 row) The reason for this is that there's a bespoke int4-to-"char" cast, which just produces the character with that code point; this is documented as the "char"(int) function. However, there are no such bespoke casts for int2 or int8, so those go via cast-via-I/O, much as if you were casting to/from text. The same happens with numeric, or indeed other types. One possible answer to this is to reclassify "char" as not being a member of the "string" category, so that the implicit-I/O-cast rule wouldn't be applied to it. I'm not sure if that'd have any downsides, but it seems like in principle it might be all right. "char" is so restricted that treating it as a general- purpose string type doesn't seem very sane. (Now that I look, I see that somebody thought that it'd be cool to assign category S to various special-purpose types like pg_mcv_list. That seems downright dangerous.) regards, tom lane
It makes sense that "char" is not part of the string family. After all, "char" != char.
In the scenarios I give, one would expect the same output for all integer width casting inputs, so you have my vote on the change you suggested.
On Thu, 2 Dec 2021 at 20:50, Tom Lane <tgl@sss.pgh.pa.us> wrote:
"Ian R. Campbell" <ian.campbell@thepathcentral.com> writes:
> I noticed that the following does not produce consistent cast results:
> select 1::int8::"char", 1::int4::"char", 1::int2::"char"
Concretely, that gives:
regression=# select 1::int8::"char", 1::int4::"char", 1::int2::"char";
char | char | char
------+------+------
1 | \x01 | 1
(1 row)
The reason for this is that there's a bespoke int4-to-"char" cast,
which just produces the character with that code point; this is
documented as the "char"(int) function. However, there are no
such bespoke casts for int2 or int8, so those go via cast-via-I/O,
much as if you were casting to/from text. The same happens with
numeric, or indeed other types.
One possible answer to this is to reclassify "char" as not being
a member of the "string" category, so that the implicit-I/O-cast
rule wouldn't be applied to it. I'm not sure if that'd have
any downsides, but it seems like in principle it might be all
right. "char" is so restricted that treating it as a general-
purpose string type doesn't seem very sane.
(Now that I look, I see that somebody thought that it'd be cool
to assign category S to various special-purpose types like
pg_mcv_list. That seems downright dangerous.)
regards, tom lane
"Ian R. Campbell" <ian.campbell@thepathcentral.com> writes: > It makes sense that "char" is not part of the string family. After all, > "char" != char. If it allowed more than one byte, maybe you could call it a "string", but it seems hard to justify without that. > In the scenarios I give, one would expect the same output for all integer > width casting inputs, so you have my vote on the change you suggested. To be clear, the change I'm thinking of would result in errors, not in silently applying the int4 cast. As a quick-n-dirty test: regression=# update pg_type set typcategory = 'x' where typname = 'char'; UPDATE 1 regression=# select 1::int8::"char", 1::int4::"char", 1::int2::"char"; ERROR: cannot cast type bigint to "char" LINE 1: select 1::int8::"char", 1::int4::"char", 1::int2::"char"; ^ regards, tom lane
I wrote: > To be clear, the change I'm thinking of would result in errors, not in > silently applying the int4 cast. As a quick-n-dirty test: > regression=# update pg_type set typcategory = 'x' where typname = 'char'; > UPDATE 1 > regression=# select 1::int8::"char", 1::int4::"char", 1::int2::"char"; > ERROR: cannot cast type bigint to "char" > LINE 1: select 1::int8::"char", 1::int4::"char", 1::int2::"char"; > ^ It turns out that changing the typcategory does have some undesirable side-effects, but I found another possible workaround. Patch posted at [1]. regards, tom lane [1] https://www.postgresql.org/message-id/2216388.1638480141%40sss.pgh.pa.us