Re: unnesting multirange data types

On Thu, Jul 15, 2021 at 6:47 PM Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> writes:
> > On 2021-Jun-19, Tom Lane wrote:
> >> I'd say let's sit on the unnest code for a little bit and see what
> >> happens.
>
> > ... So, almost a month has gone by, and we still don't have multirange
> > unnest(). Looking at the open items list, it doesn't look like we have
> > anything that would require a catversion bump. Does that mean that
> > we're going to ship pg14 without multirange unnest?
>
> > That seems pretty sad, as the usability of the feature is greatly
> > reduced. Just look at what's being suggested:
> > https://postgr.es/m/20210715121508.GA30348@depesz.com
> > To me this screams of an incomplete datatype. I far prefer a beta3
> > initdb than shipping 14GA without multirange unnest.
>
> Yeah, that seems pretty horrid. I still don't like the way the
> array casts were done, but I'd be okay with pushing the unnest
> addition.

I agree that array casts require better polymorphism and should be
considered for pg15.

+1 for just unnest().

------
Regards,
Alexander Korotkov