Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Wayne Cuddy <lists-pgsql@useunix.net> writes:
> A less bogus way of doing things is to use an EXCLUDE constraint,
> although that will restrict you to be running PG 9.0 or newer. You
> also need some way of representing the ranges as indexable objects.
> In 9.0 or 9.1, probably the best way is to use contrib/seg/ to
> represent the ranges as line segments. 9.2 will have a cleaner
> solution, ie range types.
Simple example for 9.2:
test=# create table foo (name text, id_range int4range, exclude using gist(name with =, id_range with &&));
NOTICE: CREATE TABLE / EXCLUDE will create implicit index "foo_name_id_range_excl" for table "foo"
CREATE TABLE
Time: 40,273 ms
test=*# insert into foo values ('name1', '[1,9)');
INSERT 0 1
Time: 0,660 ms
test=*# insert into foo values ('name1', '[10,19)');
INSERT 0 1
Time: 0,313 ms
test=*# insert into foo values ('name1', '[5,15)');
ERROR: conflicting key value violates exclusion constraint "foo_name_id_range_excl"
DETAIL: Key (name, id_range)=(name1, [5,15)) conflicts with existing key (name, id_range)=(name1, [1,9)).
test=!#
Great feature! Thx to all developers behing PG!
Andreas
--
Really, I'm not out to destroy Microsoft. That will just be a completely
unintentional side effect. (Linus Torvalds)
"If I was god, I would recompile penguin with --enable-fly." (unknown)
Kaufbach, Saxony, Germany, Europe. N 51.05082°, E 13.56889°
