Question

I am looking for an alternative way to check if ActivityType has any relationships as the way I am doing it now is too costly (activity_type.activities.empty?).

I thought about creating a new method such as ActivityType#linked_to_activity? which would do a query on ActivitiesActivityTypes HABTM join table. This is faster than the previously mentioned method but it is still too slow.

Any pointers are appreciated (apart from the usual “HABTM should be replaced by HM:through” non-constructive comment)

Models

class Activity < ActiveRecord::Base
  has_and_belongs_to_many :activity_types, :uniq => true
end

class ActivityType < ActiveRecord::Base
  has_and_belongs_to_many :activities
end

View

%table
  - @activity_types.each do |activity_type|
    %tr
      %td
        - if activity_type.activities.empty?
          Do something here...

Attempt at ActivityType#linked_to_activity?

class ActivityType < ActiveRecord::Base
  has_and_belongs_to_many :activities

  def linked_to_activity?
    ActivityType.find_by_sql(["SELECT * FROM activities_activity_types WHERE activity_type_id = ?", id]).empty? ? false : true
  end
end

NB:

• Running on Rails v2.3.5
• DB is MySQL with indices on foreign keys

Answer 1

I think you are on the right track. It’s a matter of preference, but I usually do it with a count:

class ActivityType < ActiveRecord::Base
  has_and_belongs_to_many :activities

  def linked_to_activity?
    self.activities.count > 0
  end
end

Edit:

As the comment below states: Using select count(*) will only retrieve one line from the DB, while select * retrieves all (unless of course you add a limit 1 at the end).

Also: I try not to litter my models with plain SQL unless it’s absolutely necessary for performance reasons.