Introduction

Ruby on Rails offers a fundamentally different way of dealing with URLS compared to most other languages. Rather than rely on the web server to control URL routing, Rails handles routing via the config/routes.rb file. This file controls every single aspect of your URLs. In this article we will cover several routing basics. Let's get started.

RESTful Routes

Before we get to routing, we must understand REST. Rails uses REST extensively for URL routing so we must have a good understanding of REST in order to understand Rails routing. REST stands for Representational State Transfer. RESTful applications typically treat the web like a resource. With REST, there are several HTTP methods that are used to represent the types of actions performed by the user and/or application.

HTTP Method What it's used for Examples
GET Retrieving a resource. Any time you navigate directly to a page or use google to navigating the page, you use the GET http method.
POST Creating a resource Older web applications used POST for everything. Many browsers only understand POST and GET. More on this below.
PUT Used to completely update a resource. Updating your user profile on some website would typically use patch with web frameworks that support it.
PATCH Used to partially update a resource. An example of this would be where you are just updating the password for your user profile.

You'll notice in the table above that I mentioned that most browsers only understand POST and GET. This isn't a problem for Rails. Rails actually has the HTTP method buried inside a hidden form element. When the page interacts with the server, the request is intercepted and reconstructed.

The Rails Routes File

Typically in a Rails routes file you specify an HTTP method followed by the page or action. For example:

config/routes.rb

get "users", to: "users#index"
# http GET method, In this instance, gets a list of users

get "users/:id", to: "users#show"
# In this instance, gets a specific user via the provided id.  For example: /users/3882

post "users", to: "users#create"
# http POST method, In this instance, used for creating a user.

put "users/:id"
# http PUT method, used for updating a resource, In this instance, updates the user.  Older versions of Rails used this for all updates.

patch "users/:id"
#  http PATCH method.  in this instance, used to partially update the user's information.

delete "users/:id"
# http DELETE method.  In this instance, used to delete a user.

All of the methods listed above use HTTP methods to tell Rails what methods to call on the controller. They are so important that Rails includes 2 special keywords to automatically create them. The first, resources tells Rails you wish to specify the above actions for a given controller):

config/routes.rb:

resources :users
# generates:
#   get "/users" -- index on your controller
#   get "/users/:id" -- show on your controller
#   get "/users/new" -- new method on your controller
#   post "/users" -- create on your controller
#   get "/users/:id/edit" -- edit method on your controller
#   put "/users/:id" -- update on your controller
#   patch "/users/:id" -- update on your controller
#   delete "/users/:id" -- destroy on your controller

The second keyword, resource (no collection), is used if you are interacting with a single resource. In this case, certain routes are omitted.

config/routes.rb:

resource :privacy_policy
# generates:
#   get "/privacy_policy" -- show on your controller
#   get "/privacy_policy/new" -- new method on your controller
#   post "/privacy_policy" -- create on your controller
#   get "/privacy_policy/edit" -- edit method on your controller
#   put "/privacy_policy" -- update on your controller
#   patch "/privacy_policy" -- update on your controller
#   delete "/privacy_policy" -- destroy on your controller

In the above block of code you will notice slight differences compared to resources. A singular resource is typically used when there is only one instance of a given resource. For instance, on a given website, there may only be one privacy policy, not many.

One important thing to note is that any resources you specify in your routes file as well as the controllers you make do NOT necessarily correspond to your models. For instance, it's entirely possible to have a controller that does a web request or one that modifies more than one model. There is nothing wrong with this.

You can quickly and easily skip certain actions for both resource and resources. To do this you use either the except keyword or only keyword. As their names imply, the except keyword tells rails to create all actions except the ones you specify, while the only keyword tells Rails to only create the given routes.

config/routes.rb:

resources :users, except: [:show] 
# generates actions for everything except for get "/users/:id"

resource :privacy_policy, only: [:show] 
# only generate the get "/privacy_policy route for the Privacy Policy.

Custom Actions

In certain cases, you may wish to have custom actions. These actions may either be on an individual resource (/users/:id) or a collection of resources (/users).

config/routes.rb:

resources "posts" do
  put :rate # generate put "/posts/:id/rate_up" 
end

resources "products"  do
  collection do
    get "most_popular" # generate  get "/products/most_popular"
  end
end

Custom actions typically come in handy to either make the URL more meaningful or to imply more functionality.

Nested Resources

Resources can also be nested:

config/routes.rb:

resources :users do
   resources :posts 
end

This would create what is called a nested resource. This lets you do things like http://yourdomain.com/users/4/posts/, which would list all of the posts for that specific user. Your controller would then receive a user_id parameter as an argument.

Aliasing and Redirects

You can also create aliases for URLs. This means you can have a user friendly url such as http://yourdomain.com/login call a specific controller and action, the new action on your sessions controller for example.

config/routes.rb:

get 'login', to: 'sessions#new'
get 'logout', to: "sessions#destroy"

You can also redirect legacy URLs to new locations using a similar mechanism. For instance:

config/routes.rb:

get 'webcatalog/', to: redirect("/products")
# redirect http://mydomain.com/webcatalog/ to http://mydomain.com/products

get "/webcatalog/product_info.php", to: redirect {|params, req| 
    begin
      id = req.params[:products_id]
      product = Product.find(id)
      "/products/#{product.slug}"
    rescue
      "/products"
    end
}

# Slightly more advanced example that uses a block to redirect from an OSCommerce page to a Rails style URL.  

Constraints

You can also specify constraints. For instance, let's say you want to constrain certain actions to a particular subdomain. You can do so with the following code.

routes.rb:

constraints subdomain: 'api' do
  resources :products do
end

Constraints can be even more advanced, for example, the following constraint checks to see if the user is using an iPhone:

routes.rb:

constraints(lambda { |req| req.env["HTTP_USER_AGENT"] =~ /iPhone/ }) do
  resources :products
end

For really advanced constraints, you can create a class. For example, the following code does the exact same thing as before:

iphone.rb:

class Iphone
  def self.matches?(request)
    request.env["HTTP_USER_AGENT"] =~ /iPhone/
  end
end
routes.rb:

constraints(Iphone) do
  resources :products
end

Namespaces

You can have custom namespaces as well. Everything in a namespace will be prefixed by the namespace. For instance:

routes.rb:

namespace :admin do
  get 'posts' => "posts#index" # http://yourapp.com/admin/posts
end

Conclusion

This isn't meant to be an exhaustive guide to routing in Ruby on Rails, but rather it's designed to help you get your feet wet without having to dig through tons of documentation. Let me know if you find this guide useful. Thanks for reading!