Andrew Gierth <andrew@tao11.riddles.org.uk> writes:
> "Tom" == Tom Lane <tgl@sss.pgh.pa.us> writes:
> Tom> Anyway, after further thought I've come up with an approach
> Tom> that's purely a syntactic transformation and so less likely to
> Tom> cause surprise: let's say that if we have TABLE() with a single
> Tom> argument, and no coldeflist either inside or outside, then we
> Tom> implicitly insert UNNEST(). Otherwise not.
> This seems ugly beyond belief.
True :-(
> If there isn't a reasonable syntax alternative to TABLE(...) for the
> multiple functions case, then frankly I think we should go ahead and
> burn compatibility with a spec feature which appears to be of negative
> value.
TBH, I'm getting close to that conclusion too. The more I look at the
spec, the more I think it must be a mistake, or else I'm somehow reading
it wrong, because it sure makes no sense for them to have invented
something that's just an alternative and less-clear syntax for a feature
they already had.
Can anyone who's following this thread check the behavior of Oracle or
DB2 to see if they interpret TABLE() the way I think the spec says?
regards, tom lane