On Thu, Dec 26, 2002 at 14:42:39 -0500,
Tom Lane <tgl@sss.pgh.pa.us> wrote:
>
> which means that views using subselects in their targetlists will not be
> flattened into the calling query in 7.3.1. This is not real desirable,
> but I see no other short-term fix.
Thanks for the explaination.
> In the particular case, your view definition seemed mighty inefficient
> anyway (it must recompute the subselects for each column retrieved from
> the view) so I think your rewrite is a good change.
I was naively expecting that the planner would notice the common subexpressions
and only compute them once.