Ok then, if it's per spec, nothing to say.
Thanks to everybody, Richard and Tom, for your time.
PS : well yes, i think it is reasonably weird. I sure don't want to try
and imagine the case you're proposing. Brain's too precious to burn. I'm
confident in you to be assured it's weirder and weirdest :-)
On Wed, Mar 16, 2005 at 10:17:17AM -0500, Tom Lane wrote:
> David Pradier <dpradier@apartia.fr> writes:
> > One can access the columns of the main query from the subquery,
> > therefore in my own query the column "id_compte" is found,
> > therefore there is no error message.
>
> > Doesn't this count as a bug ?
>
> No; it's required behavior per the SQL specification. The spec doesn't
> restrict the sub-query to reference outer columns in only some places;
> it can use them anywhere.
>
> (If you think that's weird, you should try putting the outer reference
> in an aggregate function ... it then counts as an aggregate of the outer
> query, not the inner. Also per spec.)
>
> regards, tom lane
>
> ---------------------------(end of broadcast)---------------------------
> TIP 6: Have you searched our list archives?
>
> http://archives.postgresql.org
--
dpradier@apartia.fr - tel: 01.46.47.21.33 - fax: 01.46.47.21.37
