Re: patch: function xmltable

On 12 September 2016 at 14:02, Pavel Stehule <pavel.stehule@gmail.com> wrote:
> Hi
>
> There is some opened questions - the standard (and some other databases)
> requires entering XPath expression as string literal.
>
> I am thinking so it is too strong not necessary limit - (it enforces dynamic
> query in more cases), so I allowed the expressions there.

I agree. There's no reason not to permit expressions there, and there
are many other places where we have similar extensions.

> Another questions is when these expressions should be evaluated. There are
> two possibilities - once per query, once per input row. I selected "once per
> input row mode" - it is simpler to implement it, and it is consistent with
> other "similar" generators - see the behave and related discussion to
> "array_to_string" and evaluation of separator argument. The switch to "once
> per query" should not be hard - but it can be strange for users, because
> some his volatile expression should be stable.

I would've expected once per query. There's no way the expressions can
reference the row data, so there's no reason to evaluate them each
time.

The only use case I see for evaluating them each time is - maybe -
DEFAULT. Where maybe there's a use for nextval() or other volatile
functions. But honestly, I think that's better done explicitly in a
post-pass, i.e.

select uuid_generate_v4(), x.*
from (
  xmltable(.....) x
);

in cases where that's what the user actually wants.

There's no other case I can think of where expressions as arguments to
set-returning functions are evaluated once per output row.

-- 
 Craig Ringer                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services