Hi there,
1. I have a table that looks like this:
create table zonez (p polygon);
2. I have an index that looks like this:
create index zonez__p2 on zonez using gist(p poly_ops);
3. I inserted several records in that table, it looks like:
postgres=# select * from zonez ;
p
-------------------------------------------
((1,1),(1,5),(5,5),(5,1))
((1,2),(1,5),(5,5),(5,1))
((1,3),(1,5),(5,5),(5,1))
((1,4),(1,5),(5,5),(5,1))
((1,6),(1,5),(5,5),(5,1))
((1,7),(1,5),(5,5),(5,1))
((1,7),(1,5),(5,5),(500000,1000))
((1,7),(1,5),(5,5),(52.654987,37.123789))
(8 rows)
p
-------------------------------------------
((1,1),(1,5),(5,5),(5,1))
((1,2),(1,5),(5,5),(5,1))
((1,3),(1,5),(5,5),(5,1))
((1,4),(1,5),(5,5),(5,1))
((1,6),(1,5),(5,5),(5,1))
((1,7),(1,5),(5,5),(5,1))
((1,7),(1,5),(5,5),(500000,1000))
((1,7),(1,5),(5,5),(52.654987,37.123789))
(8 rows)
4. I've "turned off" seqscan: set enable_seqscan = off;
5. I've issued vacuum analyze
6. But postgresql still doesn't want to use my index:
postgres=# explain analyze select * from zonez where '(2,2)'::point <@ p;
QUERY PLAN
---------------------------------------------------------------------------------------------------------------------
Seq Scan on zonez (cost=10000000000.00..10000000001.10 rows=1 width=101) (actual time=0.013..0.018 rows=2 loops=1)
Filter: ('(2,2)'::point <@ p)
Rows Removed by Filter: 6
Planning Time: 0.069 ms
Execution Time: 0.036 ms
(5 rows)
QUERY PLAN
---------------------------------------------------------------------------------------------------------------------
Seq Scan on zonez (cost=10000000000.00..10000000001.10 rows=1 width=101) (actual time=0.013..0.018 rows=2 loops=1)
Filter: ('(2,2)'::point <@ p)
Rows Removed by Filter: 6
Planning Time: 0.069 ms
Execution Time: 0.036 ms
(5 rows)
How come? Is it an index that should look different, or is it really more expensive than 10000000001.10?