Rails/mysql SUM distinct records - optimization

Question

Hey. How would you optimize this SQL

SELECT SUM(tmp.cost) FROM (
   SELECT DISTINCT clients.id as client, countries.credits_cost AS cost
   FROM countries
   INNER JOIN clients ON clients.country_id = countries.id
   INNER JOIN clients_groups ON clients_groups.client_id=clients.id
   WHERE clients_groups.group_id IN (1,2,3,4,5,6,7,8,9)
   GROUP BY clients.id
) AS tmp;

I'm using this example as part of my Ruby on Rails project. Note that my nested SQL (tmp) can have more then 10 milion records. You can split that in more SQLs if the performance is better. Should I add any indexes to make it quicker (i have it on IDs)?

Answer 1

You should have indexes on all the foreign keys involved in this SQL code such as: clients_groups.client_id, clients.country_id, clients_groups.group_id

However, mainly When you have that number of entries, Using SQL functions such as SUM() or COUNT() can be performance killing So I suggest you keep a cache of the cost and update it with each transaction affecting the cost.