Re: Range Types - efficiency

pgsql@j-davis.com (Jeff Davis) writes:
> On Wed, 2011-02-09 at 16:20 -0500, Chris Browne wrote: 
>> rangetest@localhost->  explain analyze select * from some_data where '[2010-01-01,2010-02-01)'::daterange @>
whensit;
>>                                                QUERY PLAN
>> ---------------------------------------------------------------------------------------------------------
>>  Seq Scan on some_data  (cost=0.00..634.00 rows=1 width=8) (actual time=1.045..111.739 rows=390 loops=1)
>>    Filter: ('[ 2010-01-01, 2010-02-01 )'::daterange @> whensit)
>>  Total runtime: 111.780 ms
>> (3 rows)
>> 
>> This, alas, reverts to a seq scan on the table, rather than restricting
>> itself to the tuples of interest.
>> 
>> I realize that, after a fashion, I'm using this backwards.  But when I'm
>> doing temporal stuff, that tends to be the pattern:
>
> Yes. The index is a btree index on a normal column, so range types can't
> exactly help with that directly -- except maybe as a rewrite like you
> say.
>
> One thing you might try is a functional index on (range(whensit)) and
> then do: where '...' @> range(whensit).
>
> Does that work for you?

That doesn't appear to actually help:

rangetest@localhost->  create index i2 on some_data (range(whensit));
CREATE INDEX
rangetest@localhost->  explain analyze select * from some_data where '[2010-01-01,2010-02-01)'::daterange @>
range(whensit);                                               QUERY PLAN
 
-------------------------------------------------------------------------------------------------------------Seq Scan
onsome_data  (cost=0.00..727.60 rows=12480 width=8) (actual time=1.030..110.542 rows=390 loops=1)  Filter: ('[
2010-01-01,2010-02-01 )'::daterange @> range(whensit))Total runtime: 110.585 ms
 
(3 rows)

In any case, I suggest that as a "couple steps down the road" thing, it
would be desirable to have that query rewrite.  Seems like a reasonable
ToDo item to consider for the future, if not in the first deployment.

Maybe that's something to add in 9.2 CommitFest #3! :-)
-- 
"There  isn't  any  reason  why  Linux  can't  be  implemented  as  an
enterprise  computing solution.   Find  out what  you've been  missing
while you've been rebooting Windows NT." - Infoworld