Re: Resolving polymorphic functions with related datatypes - Mailing list pgsql-hackers

From Simon Riggs
Subject Re: Resolving polymorphic functions with related datatypes
Date
Msg-id 1215082298.3845.769.camel@ebony.site
Whole thread Raw
In response to Re: Resolving polymorphic functions with related datatypes  ("Pavel Stehule" <pavel.stehule@gmail.com>)
List pgsql-hackers
On Thu, 2008-07-03 at 12:22 +0200, Pavel Stehule wrote:
> 2008/7/3 Simon Riggs <simon@2ndquadrant.com>:
> > I'm using the nvl() function from the orafce package. It is defined as a
> > polymorphic function so its function signature is
> >  nvl(anyelement, anyelement)
> >
> > Now if I try to use the function in this very typical way
> >  nvl(numeric_col, 0)
> >
> > we get
> >
> >  ERROR: function nvl(numeric, integer) does not exist
> >
> > The same error occurs if we have nvl(smallint, integer) etc
> >
> > This is a real shame 'cos polymorphic functions ought to be a great way
> > of saving development time and catalog space, yet they seem to fall down
> > a hole without implicit casting.
> >
> > What I'd like it to do is to recognise that the 0 should be cast
> > implicitly to another datatype within the same family. I want and expect
> >  nvl(char_column, 0)
> > to fail, but I expect the various numeric/integer types we have to play
> > nicely together without tears.
> >
> > If we can do it for indexes, can we do it for polymorphic functions also
> > when there is no matching function?
> >
> 
> +1
> 
> there is similar problem with literal constant.

as well as NULL itself, which doesn't have a type when attempting to
resolve to anyelement.

-- Simon Riggs           www.2ndQuadrant.comPostgreSQL Training, Services and Support



pgsql-hackers by date:

Previous
From: "Pavel Stehule"
Date:
Subject: Re: Resolving polymorphic functions with related datatypes
Next
From: Zdenek Kotala
Date:
Subject: Re: [WIP] patch - Collation at database level