Question

I’m no database expert, but I have enough knowledge to get myself into trouble, as is the case here. This query

SELECT DISTINCT p.* 
  FROM points p, areas a, contacts c 
 WHERE (    p.latitude > 43.6511659465 
        AND p.latitude < 43.6711659465 
        AND p.longitude > -79.4677941889 
        AND p.longitude < -79.4477941889) 
   AND p.resource_type = 'Contact' 
   AND c.user_id = 6

is extremely slow. The points table has fewer than 2000 records, but it takes about 8 seconds to execute. There are indexes on the latitude and longitude columns. Removing the clause concering the resource_type and user_id make no difference.

The latitude and longitude fields are both formatted as number(15,10) — I need the precision for some calculations.

There are many, many other queries in this project where points are compared, but no execution time problems. What’s going on?

Answer 1

Did you forget something from your actual query? It’s missing ANSI-89 joins between the three tables, giving you a cartesian product but only pulling out the POINTS records.