Re: WIP: RangeTypes - Mailing list pgsql-hackers

On Tue, Jan 11, 2011 at 01:16:47AM -0800, Jeff Davis wrote:
> Ok, I have made some progress. This is still a proof-of-concept patch,
> but the important pieces are working together.
> 
> Synopsis:
> 
>   CREATE TYPE numrange AS RANGE (SUBTYPE=numeric, 
>     SUBTYPE_CMP=numeric_cmp);
> 
>   SELECT range_eq('[1,2.2)'::numrange,'[1,2.2]');
>   SELECT range_lbound('(3.7,9]'::numrange);
>   SELECT range(6.7);
>   SELECT '-'::numrange; -- empty
>   SELECT '[1, NULL]'::numrange; -- ] will become )
>   SELECT '(INF, 3)'::numrange;
> 
> I haven't completed many of the other generic functions, because I'd
> like to make sure I'm on the right track first. The important thing
> about the functions above is that they show ANYRANGE working in
> conjunction with ANYELEMENT in various combinations, which was a
> significant part of this patch.
> 
> Here are the open items:
> 
> 1. Generic functions -- most of which are fairly obvious. However, I
> want to make sure I'm on the right track first.
> 
> 2. GiST -- I'll need a mechanism to implement the "penalty" function,
> and perhaps I'll also need additional support for the picksplit
> function. For the "penalty" function, I think I'll need to require a
> function to convert the subtype into a float, and I can use that to find
> a distance (which can be the penalty). That should also satisfy anything
> that picksplit might need.
> 
> 3. Typmod -- There is still one annoyance about typmod remaining. I need
> to treat it like an array in find_typmod_coercion_function(), and then
> create a coercion expression. Is it worth it? Would typmod on a range be
> confusing, or should I just finish this item up?

Probably not worth it for the first round.

> 4. Docs

Happy to help evenings this week :)

> 5. Tests

Same.  What do you have so far?

> 6. pg_dump -- should be pretty easy; I just want to settle some of the
> other stuff first.
> 
> 7. Right now the parse function is quite dumb. Is there some example
> code I should follow to make sure I get this right?

KISS is a fine principle.  Do you really need it smart on the first
round? :)

> 8. In order to properly support the various combinations of ANYRANGE and
> ANYELEMENT in a function definition (which are all important), we need
> to be able to determine the range type given a subtype. That means that
> each subtype can only have one associated range, which sounds somewhat
> limiting, but it can be worked around by using domains. I don't think
> this is a major limitation. Comments?

As we get a more nuanced type system, this is one of the things that
will need to get reworked, so I'd say it's better not to put too much
effort into things that a refactor of the type system
<http://wiki.postgresql.org/wiki/Refactor_Type_System> would make much
better, at least right now.

> Also related to representation:
> 
>   * Right now I always align the subtypes within the range according to
> typalign. I could avoid that by packing the bytes tightly, and then
> copying them around later. Suggestions? And what should the overall
> alignment of the range type be?

For the first cut, the simplest possible.

>   * If it's a fixed-length type, we can save the varlena header byte on
> the overall range; but we lose the ability to save space when one of the
> boundaries of the range is missing (NULL or INF), and it would
> complicate the code a little. Thoughts?

Probably not worth complicating the code at this stage.  KISS again :)

Cheers,
David.
-- 
David Fetter <david@fetter.org> http://fetter.org/
Phone: +1 415 235 3778  AIM: dfetter666  Yahoo!: dfetter
Skype: davidfetter      XMPP: david.fetter@gmail.com
iCal: webcal://www.tripit.com/feed/ical/people/david74/tripit.ics

Remember to vote!
Consider donating to Postgres: http://www.postgresql.org/about/donate