Thread: pg_prepared_statements missing column

pg_prepared_statements missing column

From
Peter Gigowski
Date:
Hello Devs,

My buddy and I were running through the 15.1 docs today and found the pg_prepared_statements view doesn't include the return_type column like it says it should.

Here is the printout from /d pg_prepared_statements
View "pg_catalog.pg_prepared_statements" Column | Type | Collation | Nullable | Default -----------------+--------------------------+-----------+----------+--------- name | text | | | statement | text | | | prepare_time | timestamp with time zone | | | parameter_types | regtype[] | | | from_sql | boolean | | | generic_plans | bigint | | | custom_plans | bigint | | |

My version:
PostgreSQL 15.1 (Debian 15.1-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit

Link to docs:

I'm not sure exactly what to make of it, but I wanted to submit a ticket.

Thanks for all your hard work!
-PG


Re: pg_prepared_statements missing column

From
Julien Rouhaud
Date:
Hi,

On Sat, Nov 12, 2022 at 11:37:34PM -0800, Peter Gigowski wrote:
>
> My buddy and I were running through the 15.1 docs today and found the
> pg_prepared_statements view doesn't include the return_type column like it
> says it should.
>
> Here is the printout from /d pg_prepared_statements
> View "pg_catalog.pg_prepared_statements" Column | Type | Collation |
> Nullable | Default
> -----------------+--------------------------+-----------+----------+---------
> name | text | | | statement | text | | | prepare_time | timestamp with time
> zone | | | parameter_types | regtype[] | | | from_sql | boolean | | |
> generic_plans | bigint | | | custom_plans | bigint | | |
>
> My version:
> PostgreSQL 15.1 (Debian 15.1-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled
> by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
>
> Link to docs:
> https://www.postgresql.org/docs/15/view-pg-prepared-statements.html
>
> I'm not sure exactly what to make of it, but I wanted to submit a ticket.

Indeed!  It's a documentation bug.  This was apparently an oversight
when packpatching 5766443695d (Bruce in Cc), as it introduced the result_types
column in the documentation in pg15 while the feature itself is only introduced
in pg16 (84ad713cf85).



Re: pg_prepared_statements missing column

From
Julien Rouhaud
Date:
On Sun, Nov 13, 2022 at 08:54:56PM +0800, Julien Rouhaud wrote:
> Hi,
> 
> On Sat, Nov 12, 2022 at 11:37:34PM -0800, Peter Gigowski wrote:
> >
> > My buddy and I were running through the 15.1 docs today and found the
> > pg_prepared_statements view doesn't include the return_type column like it
> > says it should.
> >
> > Here is the printout from /d pg_prepared_statements
> > View "pg_catalog.pg_prepared_statements" Column | Type | Collation |
> > Nullable | Default
> > -----------------+--------------------------+-----------+----------+---------
> > name | text | | | statement | text | | | prepare_time | timestamp with time
> > zone | | | parameter_types | regtype[] | | | from_sql | boolean | | |
> > generic_plans | bigint | | | custom_plans | bigint | | |
> >
> > My version:
> > PostgreSQL 15.1 (Debian 15.1-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled
> > by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
> >
> > Link to docs:
> > https://www.postgresql.org/docs/15/view-pg-prepared-statements.html
> >
> > I'm not sure exactly what to make of it, but I wanted to submit a ticket.
> 
> Indeed!  It's a documentation bug.  This was apparently an oversight
> when packpatching 5766443695d (Bruce in Cc), as it introduced the result_types
> column in the documentation in pg15 while the feature itself is only introduced
> in pg16 (84ad713cf85).

I forgot to mention that I didn't check for any other new system view
introduced in pg16, between pg15 branching and 5766443695d.  If there were they
whould likely be present in the pg15 version of the doc too.



Re: pg_prepared_statements missing column

From
Tom Lane
Date:
Julien Rouhaud <rjuju123@gmail.com> writes:
> On Sun, Nov 13, 2022 at 08:54:56PM +0800, Julien Rouhaud wrote:
>> Indeed!  It's a documentation bug.  This was apparently an oversight
>> when packpatching 5766443695d (Bruce in Cc),

Yeah, looks that way.

> I forgot to mention that I didn't check for any other new system view
> introduced in pg16, between pg15 branching and 5766443695d.  If there were they
> whould likely be present in the pg15 version of the doc too.

I did some diff'ing and couldn't identify any other discrepancies.

            regards, tom lane