2009/10/22 Tom Lane <tgl@sss.pgh.pa.us>:
> Pavel Stehule <pavel.stehule@gmail.com> writes:
>> 2009/10/22 Tom Lane <tgl@sss.pgh.pa.us>:
>>> It is not --- it's the recommended solution for emulating computed
>>> columns. The column-as-f(x) direction maybe isn't terribly exciting,
>>> but the function-as-x.col direction definitely is.
>
>> is it documented?
>
> Yes, see "SQL Functions on Composite Types" (34.4.2 as of CVS HEAD),
> down near the end of that subsection.
>
ok, thank you
Pavel
> regards, tom lane
>