Optimize the query PostgreSql-8.4 - ruby-on-rails-3

Question

I have rails controller coding as below:
#checked_contact_ids = #list.contacts.all(
:conditions => {
"contacts_lists.contact_id" => #list.contacts.map(&:id),
"contacts_lists.is_checked" => true
}
).map(&:id)
its equivalent to sql
SELECT *
FROM "contacts"
INNER JOIN "contacts_lists" ON "contacts".id = "contacts_lists".contact_id
WHERE ("contacts_lists".list_id = 67494 )
This above query takes more time to run, I want another way to run the same query with minimum time.
Is anyone knows please notice me Or is it possible? or is the above query enough for give output?
I am waiting information...................

Answer 1

I think the main problem with your original AR query is that it isn't doing any joins at all; you pull a bunch of objects out of the database via #list.contacts and then throw most of that work away to get just the IDs.
A first step would be to replace the "contacts_lists.contact_id" => #list.contacts.map(&:id) with a :joins => 'contact_lists' but you'd still be pulling a bunch of stuff out of the database, instantiating a bunch of objects, and then throwing it all away with the .map(&:id) to get just ID numbers.
You know SQL already so I'd probably go straight to SQL via a convenience method on your List model (or whatever #list is), something like this:
def checked_contact_ids
connection.execute(%Q{
SELECT contacts.id
FROM contacts
INNER JOIN contacts_lists ON contacts.id = contacts_lists.contact_id
WHERE contacts_lists.list_id = #{self.id}
AND contacts_lists.is_checked = 't'
}).map { |r| r['id'] }
end
And then, in your controller:
#checked_contact_ids = #list.checked_contact_ids
If that isn't fast enough then review your indexes on the contacts_lists table.
There's no good reason not go straight to SQL when you know exactly what data you need and you need it fast; just keep the SQL isolated inside your models and you shouldn't have any problems.