Merlin Moncure <mmoncure@gmail.com> writes:
> On Fri, Oct 17, 2014 at 9:55 AM, Marko Tiikkaja <marko@joh.to> wrote:
>> I don't know about Tom, but I didn't like that:
>> http://www.postgresql.org/message-id/5364C982.7060003@joh.to
> Hm, I didn't understand your objection:
> <quoting>
> So e.g.:
> UPDATE foo f SET f = ..;
> would resolve to the table, despite there being a column called "f"?
> That would break backwards compatibility.
> </quoting>
> That's not correct: it should work exactly as 'select' does; given a
> conflict resolve the field name, so no backwards compatibility issue.
The point is that it's fairly messy (and bug-prone) to have a syntax
where we have to make an arbitrary choice between two reasonable
interpretations.
If you look back at the whole thread Marko's above-cited message is in,
we'd considered a bunch of different possible syntaxes for this, and
none of them had much support. The "(*)" idea actually is starting to
look pretty good to me.
regards, tom lane