Re: range_agg

po 20. 1. 2020 v 1:38 odesílatel Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> napsal:

> Paul A Jungwirth <pj(at)illuminatedcomputing(dot)com> writes:
> > On Sun, Jan 19, 2020 at 12:10 AM Pavel Stehule <pavel(dot)stehule(at)gmail(dot)com>
> wrote:
> >> Now, I think so name "anymultirange" is not good. Maybe better name is
> just "multirange"
>
> > Are you sure? This function exists to be a cast to an anymultirange,
> > and I thought the convention was to name cast functions after their
> > destination type.
>
> True for casts involving concrete types, mainly because we'd like
> the identity "value::typename == typename(value)" to hold without
> too much worry about whether the latter is a plain function call
> or a special case. Not sure whether it makes as much sense for
> polymorphics, since casting to a polymorphic type is pretty silly:
> we do seem to allow you to do that, but it's a no-op.
>
> I'm a little troubled by the notion that what you're talking about
> here is not a no-op (if it were, you wouldn't need a function).
> That seems like there's something fundamentally not quite right
> either with the design or with how you're thinking about it.
>

I thinking about completeness of operations

I can to write

CREATE OR REPLACE FUNCTION fx(anyarray, anyelement)
RETURNS anyarray AS $$
SELECT $1 || ARRAY[$2]
$$ LANGUAGE sql;

I need to some functionality for moving a value to different category (it
is more generic than casting to specific type (that can hold category)

CREATE OR REPLACE FUNCTION fx(anymultirange, anyrange)
RETURNS anyrage AS $$
SELECT $1 + multirange($1)
$$ LANGUAGE sql;

is just a analogy.

Regards

Pavel

> As a comparison point, we sometimes describe subscripting as
> being a polymorphic operation like
>
> subscript(anyarray, integer) returns anyelement
>
> It would be completely unhelpful to call that anyelement().
> I feel like you might be making a similar mistake here.
>
> Alternatively, consider this: a cast from some concrete multirange type
> to anymultirange is a no-op, while any other sort of cast probably ought
> to be casting to some particular concrete multirange type. That would
> line up with the existing operations for plain ranges.
>

> regards, tom lane
>