Re: GROUP BY using tablename.* does not work if tablename has 1 column with NULL values - Mailing list pgsql-bugs

From Francisco Olarte
Subject Re: GROUP BY using tablename.* does not work if tablename has 1 column with NULL values
Date
Msg-id CA+bJJbz9OWmL_gKAPTUkHxv7HjDgcrO1aR0o-cjhhHRQSy6B7w@mail.gmail.com
Whole thread Raw
In response to RE: GROUP BY using tablename.* does not work if tablename has 1 column with NULL values  ("Narayanan Iyer" <nars@yottadb.com>)
List pgsql-bugs
Narayanan:

On Mon, 11 Oct 2021 at 14:10, Narayanan Iyer <nars@yottadb.com> wrote:
...
> You are right. My inner query was an outer join returning 3 rows. But what I did not understand was that the NULLs in
2of the 3 rows were different because 1 was a composite NULL and 1 was a ROW(NULL) and hence they showed up as 2
differentNULLs/rows even after a GROUP BY. Tom's explanation cleared it up for me. 

Perfect. null rows, row(null) and siblings are extremely tricky, I
always try to avoid them, just pointed it in case you did not run the
subquery first and counted results.

Francisco Olarte.



pgsql-bugs by date:

Previous
From: "Narayanan Iyer"
Date:
Subject: RE: GROUP BY using tablename.* does not work if tablename has 1 column with NULL values
Next
From: Emil Iggland
Date:
Subject: Re: BUG #17141: SELECT LIMIT WITH TIES FOR UPDATE SKIP LOCKED returns wrong number of rows