ActiveRecord count

Question

Here I retrieve a collection of causes from the database. Along with the information from each cause, I want to send the number of users in each cause. This way works, but if experience has told me anything there is probably a much more efficient method to do this. How can I improve this?

 def index
 @causes = Cause.near([params[:lat], params[:lng]],10)
 @causes.each do |i|
 i['count'] = i.users.count
 end
 respond_to do |format|
 format.json { render json: {:success=> 1, :data_type => 'cause',:results => @causes} }
 end
 end

Answer 1

You can perform an inner-join on the users table and then group by the causes.id.

@causes = Cause
 .near([params[:lat], params[:lng]], 10)
 .joins('users')
 .group('causes.id')
 .select('causes.*, count(users) as count')

• Emin Israfil: Thanks for the answer! I'm not trying to be a smart ass, but is this better or just different? Can you explain why? Thanks
• Trent Earl: This is more efficient because the number of queries to database is reduced to one. It also places the work onto the database which is optimized for this task through caching and indexing.
• Emin Israfil: Thanks for getting back to me. It doesn't seem to work. More info posted above :)
• Trent Earl: @user15872 What does Cause.select('causes.id,count(users) as people').joins(:users).group('causes.id').first.people output?
• Trent Earl: You say its not in the hash but AR is not returning a hash it is returning a ActiveRecord::Relation. What it is printing to the screen is not necessarily what will be in the json. What does this show? Cause.select('causes.id,count(users) as people').joins(:users).group('causes.id').as_json