On Fri, 2015-05-01 at 10:08 -0700, Tom Lane wrote:
> What really ought to happen here, IMO, is that the output columns of the
> sub-select ought to get resolved to non-unknown types while we are doing
> parse analysis of the sub-select.
So, what would happen for something like:
select u+i from (select '1' as u, '2'::int as i) s;
?
I can see two possibilities:
1. Resolve "u" from the subselect as text, and later fail to find a
match for +(text,int).
2. Resolve +(unknown, int) to +(int, int) first, then inform the
subselect that it's looking for an int (in the same way that you propose
the insert pass down some context).
I don't think the second one really makes sense though. For example:
select u+i, u||'suffix'::text from (select '1' as u, '2'::int as i) s;
In that case, "+" would be resolved to +(int, int) and || would be
resolved to ||(text, text). But "u" from the subselect can't be both an
int and text.
Then again, we probably want to fail a query like that anyway. So maybe
it does make sense as long as we can figure out a single type for "u",
and we fail otherwise.
Regards,
Jeff Davis
