"David G. Johnston" <david.g.johnston@gmail.com> writes:
> On Wed, Mar 23, 2022 at 10:10 AM Viliam Ďurina <viliam.durina@gmail.com>
> wrote:
>> Now I'm surprised that a set-returning function is even allowed in SELECT
>> clause where the values have to be scalar.
> AFAIK the lateral construct, which is required to avoid doing just this, is
> a relatively recent invention for SQL. I infer from that fact that the
> ability to execute a set-returning function in the select clause has always
> been allowed.
I believe that Postgres' handling of that is actually a hangover
from Berkeley's PostQUEL language. Dunno what the SQL standard has
to say on the subject --- but it wouldn't surprise me if they don't
allow it.
regards, tom lane