Re: Index use in BETWEEN statement...

On 9/26/05 11:26 AM, "Cristian Prieto" <cristian(at)clickdiario(dot)com> wrote:

>
> Hello pals, I have the following table in Postgresql 8.0.1
>
> Mydb# \d geoip_block
> Table "public.geoip_block"
> Column | Type | Modifiers
> -------------+--------+-----------
> locid | bigint |
> start_block | inet |
> end_block | inet |
>
> mydb# explain analyze select locid from geoip_block where
> '216.230.158.50'::inet between start_block and end_block;
> QUERY PLAN
> ----------------------------------------------------------------------------
> -------------------------------------------
> Seq Scan on geoip_block (cost=0.00..142772.86 rows=709688 width=8) (actual
> time=14045.384..14706.927 rows=1 loops=1)
> Filter: (('216.230.158.50'::inet >= start_block) AND
> ('216.230.158.50'::inet <= end_block))
> Total runtime: 14707.038 ms
>
> Ok, now I decided to create a index to "speed" a little the query
>
> Mydb# create index idx_ipblocks on geoip_block(start_block, end_block);
> CREATE INDEX
>
> clickad=# explain analyze select locid from geoip_block where
> '216.230.158.50'::inet between start_block and end_block;
> QUERY PLAN
> ----------------------------------------------------------------------------
> ------------------------------------------
> Seq Scan on geoip_block (cost=0.00..78033.96 rows=230141 width=8) (actual
> time=12107.919..12610.199 rows=1 loops=1)
> Filter: (('216.230.158.50'::inet >= start_block) AND
> ('216.230.158.50'::inet <= end_block))
> Total runtime: 12610.329 ms
> (3 rows)
>
> I guess the planner is doing a sequential scan in the table, why not use the
> compound index? Do you have any idea in how to speed up this query?

Did you vacuum analyze the table after creating the index?

Sean