Thread: [GENERAL] RETURNS SETOF RECORD with 1 column

Hi all.

Please, see this fragment:

DROP TABLE IF EXISTS atable;
CREATE TABLE atable (
 atext TEXT,
 anint INT
);

TRUNCATE atable;
INSERT INTO atable VALUES
 ( 'one',1 ),
 ( 'two',2 ),
 ( 'three',3 ),
 ( 'four',4 )
;

CREATE OR REPLACE FUNCTION afun1( OUT ot TEXT )
RETURNS SETOF RECORD
LANGUAGE PLPGSQL
AS $l0$
BEGIN
 RETURN QUERY SELECT atext FROM atable;
END;
$l0$;

CREATE OR REPLACE FUNCTION afun2( OUT ot TEXT, OUT oi INT )
RETURNS SETOF RECORD
LANGUAGE PLPGSQL
AS $l0$
BEGIN
 RETURN QUERY SELECT atext,anint FROM atable;
END;
$l0$;

I am not allowed to create afun1() but can create and use afun2().
The error message for afun1() reads:

ERROR:  function result type must be text because of OUT parameters

The idea I understand is that a RECORD must have no less than TWO columns.
Of course it's not a problem to declare "RETURNS SETOF TEXT", but why
forcing me to "deviate" from a very general construct for a 1 column
case?

Is this a bug/enhancement or am I missing something?

TALIA!

--
Vincenzo Romano - NotOrAnd.IT
Information Technologies
--
NON QVIETIS MARIBVS NAVTA PERITVS

Vincenzo Romano <vincenzo.romano@notorand.it> writes:
> CREATE OR REPLACE FUNCTION afun1( OUT ot TEXT )
> RETURNS SETOF RECORD

> The error message for afun1() reads:
> ERROR:  function result type must be text because of OUT parameters

Indeed.  It's trying to protect you from simple typos.

If you want a consistent syntax I'd suggest

CREATE OR REPLACE FUNCTION afun1() RETURNS TABLE (ot text) ...

It's still really "setof text" under the hood.

            regards, tom lane

2017-07-27 16:03 GMT+02:00 Tom Lane <tgl@sss.pgh.pa.us>:

Thanks a lot for your reply with valuable details.

> Vincenzo Romano <vincenzo.romano@notorand.it> writes:
>> CREATE OR REPLACE FUNCTION afun1( OUT ot TEXT )
>> RETURNS SETOF RECORD
>
>> The error message for afun1() reads:
>> ERROR:  function result type must be text because of OUT parameters
>
> Indeed.  It's trying to protect you from simple typos.

Which typo?
It looks it's preventing me from consistently using another simple
syntax for FRSes.

CREATE OR REPLACE FUNCTION afun1( OUT ot TEXT )
RETURNS SETOF TEXT ...

CREATE OR REPLACE FUNCTION afun2( OUT ot TEXT, OUT it INT )
RETURNS SETOF RECORD ...

CREATE OR REPLACE FUNCTION afun3( OUT ot TEXT, OUT it INT, OUT on NUMERIC )
RETURNS SETOF RECORD ...

CREATE OR REPLACE FUNCTION afun4( OUT ot TEXT, OUT it INT, OUT on
NUMERIC, OUT od DATE )
RETURNS SETOF RECORD ...

This means that RETURNS SETOF RECORD syntax is only available from 2 columns on.
While RETURNS TABLE() syntax is always available, *also* for 1 column.
I thought that "RETURNS SETOF RECORD" was meant for FRSes just like
RETURN TABLE() is.
Why not, then?

> If you want a consistent syntax I'd suggest
>
> CREATE OR REPLACE FUNCTION afun1() RETURNS TABLE (ot text) ...

Indeed.
The main difference is that with RETURNS SETOF RECORD I still get the
"usual"(tm) function argument list in the usual place: between two
parentheses.
It's a matter of style. And a consistent one.
But I still don't get the point for not having it for a single column.
Surely it's my fault.

> It's still really "setof text" under the hood.
>
>                         regards, tom lane

--
Vincenzo Romano - NotOrAnd.IT
Information Technologies
--
NON QVIETIS MARIBVS NAVTA PERITVS

The main difference is that with RETURNS SETOF RECORD I still get the
"usual"(tm) function argument list in the usual place: between two
parentheses.
It's a matter of style. And a consistent one.
But I still don't get the point for not having it for a single column.

On Thursday, July 27, 2017, Vincenzo Romano <vincenzo.romano@notorand.it> wrote:

The main difference is that with RETURNS SETOF RECORD I still get the
"usual"(tm) function argument list in the usual place: between two
parentheses.
It's a matter of style. And a consistent one.
But I still don't get the point for not having it for a single column.

Docs say: " When there are OUT or INOUT parameters, the RETURNS clause can be omitted. ". Sounds like you should you do just that.

2017-07-28 8:36 GMT+02:00 David G. Johnston <david.g.johnston@gmail.com>:
> On Thursday, July 27, 2017, David G. Johnston <david.g.johnston@gmail.com>
> wrote:
>>
>> On Thursday, July 27, 2017, Vincenzo Romano <vincenzo.romano@notorand.it>
>> wrote:
>>>
>>> The main difference is that with RETURNS SETOF RECORD I still get the
>>> "usual"(tm) function argument list in the usual place: between two
>>> parentheses.
>>> It's a matter of style. And a consistent one.
>>> But I still don't get the point for not having it for a single column.
>>>
>>
>> Docs say: " When there are OUT or INOUT parameters, the RETURNS clause can
>> be omitted. ". Sounds like you should you do just that.
>>
>
> Except you'd have no where to put the "setof" modifier...So, yeah, you
> probably aren't going to personal style preference catered to here.
>
> David J.

Thanks David.

I have found the actual documentation of the inconsistent syntax.

It's here
https://www.postgresql.org/docs/9.6/static/sql-createfunction.html
at the description for the parameter "rettype":

[QUOTE]
When there are OUT or INOUT parameters, the RETURNS clause can be omitted.
If present, it must agree with the result type implied by the output parameters:
RECORD if there are **multiple output parameters**, or the same type
as the single output parameter.
The SETOF modifier indicates that the function will return a set of
items, rather than a single item.
The type of a column is referenced by writing table_name.column_name%TYPE.
[/QUOTE]

That single predicate, "multiple output parameters", is creating the
(useless?) special case for a single column output.
I would like to understand the typo protection mentioned by Tom earlier:
I need to understand the reason for creating that special case.

--
Vincenzo Romano - NotOrAnd.IT
Information Technologies
--
NON QVIETIS MARIBVS NAVTA PERITVS

Vincenzo Romano <vincenzo.romano@notorand.it> writes:
> I would like to understand the typo protection mentioned by Tom earlier:
> I need to understand the reason for creating that special case.

Well, case A:

create function foo(out x int4) returns setof int8 ...

This is indubitably a typo.

Case B:

create function foo(out x record) returns setof record ...

Now what?  Is the user expecting us to wrap x in an additional
layer of composite, or not?  Our current assumption is "not",
but it would be pretty inconsistent to do that if x did get
wrapped as long as it were any other type.

Yes, we could have resolved that ambiguity in one direction or the
other and then said that "returns record" or "returns setof record"
is OK regardless of the number of OUT parameters, but we didn't.
Considering that the SQL-standard syntax for this is TABLE(), and that
hasn't got these issues in the first place, I don't feel any need to
revisit the question.

            regards, tom lane