Re: Re: new type proposal

On Tue, 6 Feb 2001, Joseph Shraibman wrote:

> Alex Pilosov wrote:
> >
> > On Tue, 6 Feb 2001, Dan Wilson wrote:
> >
> > > What would this do that would be non-standard?  Does the SERIAL datatype add
> > > something that is not standard?  No... it just allows for an easy way to
> > > implement something that is standard.  The SERIAL "type" isn't really a
> > > datatype, it's just a keyword that allows you to automatically specify an
> > > int4 column with a related sequence and default.  I don't see why the same
> > > thing couldn't be done with TIMESTAMP!
> > Such way the madnesssH^H^H^Hmysql lies ;)
> >
> > I firmly believe that people who need that feature should implement it
> > themselves via triggers, and rest of us shouldn't suffer from the code
> > bloat resulting to support this.
>
> I noticed that people are ignoring the time created part of my
> proposal.  How can a read only field be implemented?  A trigger that
> causes and error if that field is updated?

That'd be one way of doing it, if the value is modified to
something distinct raise an exception...

'begin
 if (NEW.b!=OLD.b) then
  RAISE EXCEPTION ''...''';
 end if;
 return NEW;
end;'