Tom Lane escreveu:
> Bruce Momjian <bruce@momjian.us> writes:
>> Yea, I thought we were going to do this:
>
>> Please find enclosed one way to handle it, this being prepending
>> WINDOW to the result types in \df.
>
>> but I don't see this behavior in CVS.
>
> IIRC, my original proposal involved adding something to the argument
> list --- it seems more natural to regard window-ness as having something
> to do with the arguments than the result. But that was shot down on the
> grounds of not fitting in well unless we wanted to add more decoration,
> like parens around the regular argument list.
>
> Another idea was to add a new column to the \df output to mark
> window-ness. Which, as I recall, *nobody* liked. But maybe if we
> only did it for \df+ it would be more tolerable?
>
Adding another column to \df+ is not a good idea; there are already too much
columns. Window functions are special functions (they even have an different
syntax and separate section in docs) and are not less special than aggregate
functions. So +1 to add \dw for them.
-- Euler Taveira de Oliveira http://www.timbira.com/
