Re: generate_series regression 9.6->10

Thanks Tom. This showed up in a regression test of ours that built the test
data using generate_series, so it's not a critical production issue or
anything, just a surprise change in behaviour.

P.

On Wed, May 24, 2017 at 10:28 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:

> Paul Ramsey <pramsey(at)cleverelephant(dot)ca> writes:
> > The behaviour of generate_series seems to have changed a little, at least
> > in conjunction w/ CTEs.
>
> What's changed is the behavior of multiple SRFs in a SELECT's targetlist,
> cf
>
> https://git.postgresql.org/gitweb/?p=postgresql.git&a=commitdiff&h=
> 69f4b9c85f168ae006929eec44fc44d569e846b9
>
> specifically this comment:
>
> While moving SRF evaluation to ProjectSet would allow to retain the old
> "least common multiple" behavior when multiple SRFs are present in one
> targetlist (i.e. continue returning rows until all SRFs are at the
> end of
> their input at the same time), we decided to instead only return rows
> till
> all SRFs are exhausted, returning NULL for already exhausted ones. We
> deemed the previous behavior to be too confusing, unexpected and
> actually
> not particularly useful.
>
> I see the current v10 release notes have failed miserably at documenting
> this :-(. Will try to improve that.
>
> regards, tom lane
>