Re: Retrieving multiple columns from a subquery
| От | Tom Lane |
|---|---|
| Тема | Re: Retrieving multiple columns from a subquery |
| Дата | |
| Msg-id | 11172.1337053018@sss.pgh.pa.us обсуждение исходный текст |
| Ответ на | Re: Retrieving multiple columns from a subquery (Chris Hanks <christopher.m.hanks@gmail.com>) |
| Ответы |
Re: Retrieving multiple columns from a subquery
|
| Список | pgsql-general |
Chris Hanks <christopher.m.hanks@gmail.com> writes:
> Nothing? Are subqueries just not meant to be used this way?
The SQL standard says not ;-).
You could approximate it like this:
select ..., (select row(x,y,z) from ...), ... from ...;
as long as you don't mind pulling the composite-value output syntax
apart. This avoids the single-output-column syntactic restriction
by cramming all the values into one column.
[ thinks for a bit... ] It seems like you ought to be able to get PG
to pull the composite values apart again, with something like
select ..., (x).*, ... from
(select ..., (select row(x,y,z) from ...) as x, ...
from ... offset 0) ss;
but when I try this I get
ERROR: record type has not been registered
That's a bug, probably, but dunno how hard to fix. In the meantime you
could work around it by casting the row() expression to a named
composite type; which might be a good idea anyway since there's no other
obvious way to control the column names that will be exposed by the
(x).* expansion.
regards, tom lane
В списке pgsql-general по дате отправления: