Re: BUG #5234: ALTER TABLE ... RENAME COLUMN change view definition incorrectly

On Thu, Dec 10, 2009 at 11:54 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> On Thu, Dec 10, 2009 at 1:46 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> My reading of the spec is that USING (and therefore NATURAL) is defined
>>> to join identically named columns.  Therefore, renaming one of the input
>>> columns as the OP did *should* indeed *must* break the view.  The problem
>>> is not how to make it work, it's how to give an error message that
>>> doesn't look like an internal failure.
>
>> That seems ugly and unnecessary.  I think we might be able to define
>> ourselves out of this problem.  We don't guarantee (and have never
>> guaranteed) that selecting from a stored view will produce the same
>> results as re-executing the original query.  For example, * refers the
>> list of columns at definition-time, not execution-time,
>
> Um, aren't you contradicting yourself there?

Not that I can see.

> The problem with USING is that it is not merely a join condition but
> affects the set of columns emitted by the join.  It can't be converted
> to a simple ON without changing the semantics, and I don't believe we
> should try.

The OP seemed to be pretty clear on what the semantics should be, and
I'm not confused either. Why are you? :-)

...Robert