Re: DEFAULT of domain ignored in plpgsql (8.4.1) - Mailing list pgsql-hackers

From Florian G. Pflug
Subject Re: DEFAULT of domain ignored in plpgsql (8.4.1)
Date
Msg-id 4B085C01.9040009@phlo.org
Whole thread Raw
In response to Re: DEFAULT of domain ignored in plpgsql (8.4.1)  (Gurjeet Singh <singh.gurjeet@gmail.com>)
List pgsql-hackers
Gurjeet Singh wrote:
> On Sat, Nov 21, 2009 at 7:26 AM, Josh Berkus <josh@agliodbs.com 
> <mailto:josh@agliodbs.com>> wrote: However, there are some other
> issues to be resolved:
> 
> (1) what should be the interaction of DEFAULT parameters and domains 
> with defaults?
> 
> The function's DEFAULT parameter should take precedence over the
> default of the domain.

I think Josh was pondering whether

create domain myint as int default 0;
create function f(i myint) ...;

should behave like

create function f(i myint default 0) ...;

and hence call f(0) if you do "select f();", or instead
raise an error because no f with zero parameters is defined (as it does
now).

I'd say no, because "no default" should be treated the same as "default
null", so for consistency we'd then have to also support

create function g(i int) ...;
select g();

And of course throw an error if there was another function defined as
create function g() ...;

This way leads to madness...

If one really wanted to do that, there'd have to be an OPTIONAL clause
for function parameters that works like DEFAULT, but doesn't take a
default value and instead uses the type's default (NULL except for
domains with DEFAULT clause). But I wouldn't got that far, personally...

best regards,
Florian Pflug



pgsql-hackers by date:

Previous
From: "Florian G. Pflug"
Date:
Subject: Re: DEFAULT of domain ignored in plpgsql (8.4.1)
Next
From: Tom Lane
Date:
Subject: Ignoring white space in regression tests really a good idea?