This little snippet is great! The only problem I see is that the enums must
be consistent across all modules.
What about loading a variable with a "default" value? Then it could be
adjusted to 'play'.
On 10/27/05, Jim C. Nasby <jnasby(at)pervasive(dot)com> wrote:
> This is cool; it's something people can use today if nothing else.
> Long-term, is it practical to have the enums compiled in? ISTM that's
> not very workable, but I'm completely guessing. The other issue is that
> this version makes it very difficult to change what's in the enum (not
> that that's at all easy with MySQL...)
> On Thu, Oct 27, 2005 at 03:47:53PM -0400, Andrew Dunstan wrote:
> > If people would like to play, I have created a little kit to help in
> > creating first class enum types in a few seconds. It works something
> > like this:
> > make TYPENAME=rainbow ENUMS=' "red", "orange", "yellow", "green",
> > "blue", "indigo", "violet" '
> > make TYPENAME=rainbow install
> > psql -f /path/to/contrib/rainbow-install.sql yourdb
> > and you are done. Now you can do:
> > create table foo( r rainbow);
> > insert into foo values('red');
> > select 'red'::rainbow < 'green'::rainbow; <-- yields true
> > select rainbow_order('yellow'); <-- yields 2
> > The kit is at http://developer.postgresql.org/~adunstan/enumkit.tgz
> > Needs 8.0 or later, since it use PGXS.
> > Maximum number of values is 32767 - but if you use that many you're
> > insane anyway :-)
> > I did this as part of thinking about how we might do enums properly. AS
> > Chris KL recently noted - it is very often asked for. So this is not the
> > end of the road, just a tiny step at the beginning.
> > cheers
> > andrew
In response to
- Re: enums at 2005-10-27 20:46:03 from Jim C. Nasby
- Re: enums at 2005-10-27 21:26:13 from Jim C. Nasby
- Re: enums at 2005-10-28 02:20:51 from Andrew Dunstan
pgsql-hackers by date
|Next:||From: Grzegorz Piotr Jaskiewicz||Date: 2005-10-27 21:15:53|
|Subject: _penalty gist method invoked with one key NULL|
|Previous:||From: Jim C. Nasby||Date: 2005-10-27 20:46:03|
|Subject: Re: enums|