I want to compare the performance of some queries so I created those tables:
CREATE TABLE reservation1 (room int, name varchar(255), during daterange);
CREATE TABLE reservation2 (room int, name varchar(255), start_date date, end_date date);
then:CREATE TABLE reservation1 (room int, name varchar(255), during daterange);
CREATE TABLE reservation2 (room int, name varchar(255), start_date date, end_date date);
INSERT INTO public.reservation1(name, room, during)
select concat('room ', a), a, '[2010-01-01, 2010-04-02)' from generate_series(1,10000000) as a;
INSERT INTO public.reservation2(name, room, start_date, end_date)
select concat('room ', a), a, '2010-01-01', '2010-04-02' from generate_series(1,10000000) as a;
I create an index for during column:
CREATE INDEX reservation1_idx ON reservation1 USING GIST (during);
CREATE INDEX reservation1_idx ON reservation1 USING GIST (during);
I'm using the operator contains (@>) and overlaps (&&):
select * from reservation1
where during @> '[2010-02-15,2010-03-02)';
select * from reservation1
where during && '[2010-02-15,2010-03-02)';
select * from reservation1
where during @> '[2010-02-15,2010-03-02)';
select * from reservation1
where during && '[2010-02-15,2010-03-02)';
And I got such result time: 25s
However when I use this query:
select * from reservation2
where ('2010-02-15' between start_date and end_date)
and ('2010-03-02' between start_date and end_date);
select * from reservation2
where ('2010-02-15' between start_date and end_date)
and ('2010-03-02' between start_date and end_date);
The result time was: 9s.
I understand that the index does not have any effect when the amount of fetched data was huge and the query planner used the seq scan method. I want to know if it is not recommended to use rang types and operator to get good performance or should I add something to the queries to be faster?
