Re: Function proposal to find the type of a datum

From: Kate F <kate(at)cats(dot)meow(dot)at>
To: pgsql-hackers(at)postgresql(dot)org
Subject: Re: Function proposal to find the type of a datum
Date: 2007-02-02 07:31:13
Message-ID: 20070202073113.GA390@cats.meow.at
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Fri, Feb/ 2/07 02:17:51AM -0500, Tom Lane wrote:
> "Pavel Stehule" <pavel(dot)stehule(at)hotmail(dot)com> writes:
> > you can identify type via operator OF.
>
> > IF a IS OF INTEGER THEN
> > RAISE NOTICE 'Parametr a is numeric';
>
> Yeah, that is the SQL-standard syntax, but I think our implementation
> doesn't work the way Kate would like: if "a" is an ANYELEMENT function
> parameter then (I think) only IS OF ANYELEMENT would succeed. Would it
> conform to the SQL spec for the test to "look through" ANYELEMENT to the
> actual parameter type? Would it be reasonably implementable to do that?

Well, actually, I discussed this on IRC with Pavel, and for the above I
have syntax error at or near "INTEGER". My guess was that the grammar
changed as an unintentional side-effect of some other change. He corrected
his suggestion: OF takes a list of types; it needs parentheses:

IF a IS OF (INTEGER) THEN

This operator would look wonderful for my particular needs, were it not
that it has completely different semantics! It appears to test to see
what a datum actually *is* (in my case TEXT), rather than what it may
be parsed as (in my case, INTEGER).

Meanwhile, I still think the function David proposed is a worthy
addition (and I still have a user-case for it!), as using just the OF
operator for something similar, one would have to explictly test
against every type required.

Thank you,

--
Kate

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Tom Lane 2007-02-02 07:41:15 Re: Function proposal to find the type of a datum
Previous Message Tom Lane 2007-02-02 07:17:51 Re: Function proposal to find the type of a datum