On Sat, Apr 11, 2009 at 10:32:14AM -0400, Tom Lane wrote:
> David Fetter <david@fetter.org> writes:
> > On Sat, Apr 11, 2009 at 08:52:31AM -0400, Robert Haas wrote:
> >> We're up to at least four different categories of functions that
> >> people think might require special treatment: window, trigger,
> >> I/O, everything else.
>
> > The current psql has \da and \df, the latter of which now includes
> > I/O functions. I contend that windowing functions are different
> > enough that they require a separate category.
>
> I think the fact that aggregates have a separate command is somewhat
> historical. However, the fact remains that at the SQL level there
> is CREATE/DROP/etc AGGREGATE and CREATE/DROP/etc FUNCTION, and
> nothing else. If we don't hang psql's hat on that same hook then we
> are going to confuse users --- not to mention that this thread will
> never reach a resolution because there will be too many
> alternatives.
The "do nothing" solution is unacceptable because windowing functions
behave in a way that's essentially different, from the user's
perspective, from other functions including aggregates. Speaking of
aggregates, they should probably show up in the windowing functions
section too, as they behave differently there. For example, a sum()
over a window with ordering is a *running* sum (to the extent that the
ORDER BY clause causes unique values), a completely different behavior
from its normal aggregate/non-ordered windowing behavior.
Cheers,
David.
--
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778 AIM: dfetter666 Yahoo!: dfetter
Skype: davidfetter XMPP: david.fetter@gmail.com
Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate
