Scaffolding ActiveRecord: two columns of the same data type

Another basic Rails question:

I have a database table that needs to contain references to exactly two different records of a specific data type.

Hypothetical example: I'm making a video game database. I have a table for "Companies." I want to have exactly one developer and exactly one publisher for each "Videogame" entry.

I know that if I want to have one company, I can just do something like:

script/generate Videogame company:references

But I need to have both companies. I'd rather not use a join table, as there can only be exactly two of the given data type, and I need them to be distinct.

It seems like the answer should be pretty obvious, but I can't find it anywhere on the Internet.

Answers


Just to tidy things up a bit, in your migration you can now also do:

create_table :videogames do |t|
  t.belongs_to :developer
  t.belongs_to :publisher
end

And since you're calling the keys developer_id and publisher_id, the model should probably be:

belongs_to :developer, :class_name => "Company"
belongs_to :publisher, :class_name => "Company"

It's not a major problem, but I find that as the number of associations with extra arguments get added, the less clear things become, so it's best to stick to the defaults whenever possible.


I have no idea how to do this with script/generate.

The underlying idea is easier to show without using script/generate anyway. You want two fields in your videogames table/model that hold the foreign keys to the companies table/model.

I'll show you what I think the code would look like, but I haven't tested it, so I could be wrong.

Your migration file has:

create_table :videogames do |t|
  # all your other fields
  t.int :developer_id
  t.int :publisher_id
end

Then in your model:

belongs_to :developer, class_name: "Company", foreign_key: "developer_id"
belongs_to :publisher, class_name: "Company", foreign_key: "publisher_id"

You also mention wanting the two companies to be distinct, which you could handle in a validation in the model that checks that developer_id != publisher_id.


If there are any methods or validation you want specific to a certain company type, you could sub class the company model. This employs a technique called single table inheritance. For more information check out this article: http://wiki.rubyonrails.org/rails/pages/singletableinheritance

You would then have:

#db/migrate/###_create_companies
class CreateCompanies < ActiveRecord::Migration
  def self.up
    create_table :companies do |t|
      t.string :type  # required so rails know what type of company a record is
      t.timestamps
    end
  end

  def self.down
    drop_table :companies
  end
end

#db/migrate/###_create_videogames
class CreateVideogames < ActiveRecord::Migration
  create_table :videogames do |t|
    t.belongs_to :developer
    t.belongs_to :publisher
  end    

  def self.down
    drop_table :videogames
  end
end

#app/models/company.rb
class Company < ActiveRecord::Base 
  has_many :videogames
  common validations and methods
end

#app/models/developer.rb
class Developer < Company
  developer specific code
end

#app/models/publisher.rb
class Publisher < Company
  publisher specific code
end

#app/models/videogame.rb
class Videogame < ActiveRecord::Base 
  belongs_to :developer, :publisher
end

As a result, you would have Company, Developer and Publisher models to use.

 Company.find(:all)
 Developer.find(:all)
 Publisher.find(:all)

Need Your Help

Converting Django QuerySet to pandas DataFrame

python django pandas

I am going to convert a Django QuerySet to a pandas DataFrame as follows:

A Python script that activates the virtualenv and then runs another Python script?

python windows shell command virtualenv

On Windows&nbsp;Vista, I need a script that starts the activate (to activate the virtualenv) script in: