PG-incompatibility: parentheses after GROUP BY shall not be row constructor

[xiangjinwu]

When used in SELECT list, a pair of parentheses is treated as simplified version of row constructor:

test=# create table t (id int, cat varchar, at timestamptz);
CREATE TABLE
test=# insert into tt values (1, 'foo', now()), (2, 'foo', now());
INSERT 0 2

test=# select id, cat from t;
 id | cat 
----+-----
  1 | foo
  2 | foo
(2 rows)

test=# select (id, cat) from t;
   row   
---------
 (1,foo)
 (2,foo)
(2 rows)

test=# select row(id, cat) from t;
   row   
---------
 (1,foo)
 (2,foo)
(2 rows)

However, this shall not be the case after GROUP BY:

test=# select id from t group by 3+2, (id, cat);
 id 
----
  2
  1
(2 rows)

test=# select id from t group by 3+2, row(id, cat);
ERROR:  column "t.id" must appear in the GROUP BY clause or be used in an aggregate function
LINE 1: select id from t group by 3+2, row(id, cat);
               ^

As of now, RisingWave reports the same error in both cases, treating parentheses same as row constructor.

[github-actions]

This issue has been open for 60 days with no activity. Could you please update the status? Feel free to continue discussion or close as not planned.