Notes posted to Ruby on Rails

RSS feed
March 27, 2012
4 thanks

Makes it possible to use a scope through an association

This is a very useful method if you want to to use a scope through an association:

class Book < ActiveRecord::Base
  scope :available, where(:available => true)

class Author < ActiveRecord::Base
  has_many :books
  scope :with_available_books, joins(:books).merge(Book.available)

# Return all authors with at least one available book:

See http://asciicasts.com/episodes/215-advanced-queries-in-rails-3 for more info.

March 27, 2012
6 thanks


If you want to override previously set order (even through default_scope), use reorder() instead.


User.order('id ASC').reorder('name DESC')

would ignore ordering by id completely

March 27, 2012 - (>= v3.2.1)
0 thanks

Example to auto download


op = Operation.find(params[:id])
fname = "operation_#{op.id}_#{DateTime.now.to_i}.csv"
send_data op.export(params[:url_type]), 
  :type => 'text/csv; charset=iso-8859-1; header=present',
  :disposition => "attachment; filename=#{fname}.csv"


def export(url_type)
  csv_data = CSV.generate do |csv|
    csv << self.header_columns # simple array ["id","name"]
    url_items = @operation.url_items.where(:url_type => url_type)
    url_items.each do |url_item|
      csv << self.process_row(url_item)  # simple array [1,"bob"]
  return csv_data
March 26, 2012 - (v2.3.2 - v3.1.0)
1 thank

Use exist scopes on default_scope - pay attention

To use exists scopes on default_scope , you can use something like:

class Article < ActiveRecord::Base
  scope :active, proc {
    where("expires_at IS NULL or expires_at > '#{Time.now}'")

  scope :by_newest, order("created_at DESC")

  default_scope by_newest

But, if you would add a filter, and it require a lazy evaluate, use block on default_scope declaration, like:

default_scope { active.by_newest }
March 22, 2012 - (>= v3.1.0)
2 thanks

Deprecation in 3.1+

In Rails 3.1 and higher, just use ruby’s SecureRandom, e.g.




March 15, 2012 - (>= v3.2.1)
1 thank

Deprecated proxy_owner

Just change your


calls to


Found it here: http://mileszs.com/deprecation-warnings-for-proxyowner-in-rails

March 10, 2012
3 thanks

flash messages

In rails 3.1 the following does not work for me

redirect_to { :action=>'atom' }, :alert => "Something serious happened" 

Instead, you need to use the following syntax (wrap with parens)

redirect_to({ :action=>'atom' }, :alert => "Something serious happened")
March 5, 2012
0 thanks

Example of conditions using


validates :number, :presence => { :if => :quota_file? }

def self.quota_file?
March 5, 2012
0 thanks

won't refresh updated_at

This will not cause :updated_at column to refresh, while ActiveRecord::Base#increment! would.

February 20, 2012
0 thanks

How to user method 'from'


# => "SELECT `users`.* FROM posts"
February 20, 2012
0 thanks

How to user method from'


# => "SELECT `users`.* FROM posts"
February 15, 2012 - (<= v2.3.8)
0 thanks

@ssoroka and @drova and future readers

I guess these two have already found a solution, but future readers might have not. index and references do not map perfectly

change_table :foo do |t|
  t.references :bar
  t.index      :bar_id

references gets the model name while index gets the column name.

February 15, 2012 - (<= v2.3.8)
0 thanks

Not working in create_table

When using the index method within a create_table statement, it does not have any side effect - at least not in MySQL.

create_table :comment do |t|
  t.belongs_to :post


  # not working inside create_table !
  t.index :post_id

It is working properly in change_table though

change_table :comment do |t|
  t.belongs_to :user

  # this works inside change_table
  t.index :user_id

Unfortunately this flaw is not reported in any way. The index is just not created.

I have only tested this with the mysql2 driver in Rails 2.3.x. I’m not sure, if this happens in other versions/adapters as well.

February 10, 2012 - (>= v3.1.0)
0 thanks

:defaults no longer work

I’m afraid that :defaults option doesn’t work anymore.

<%= javascript_include_tag :defaults %>

it loads “defaults.js”

Instead, to load all .js files now use

<%= javascript_include_tag "application" %>
February 9, 2012 - (v1.2.0 - v3.1.0)
1 thank

:only, :except and passing in multiple parameters

To specify that the filter should be applied to or excluded from given controller actions, use the :only and :except parameters. To pass in multiple controller actions use an array:

after_filter :authorize, :except => [:index, :show]
after_filter :authorize, :only => :delete

Stolen from: http://apidock.com/rails/ActionController/Filters/ClassMethods/before_filter

February 9, 2012 - (v2.3.2 - v3.1.0)
0 thanks


This is inverse operation of select. If block return false add item to array.

[1,2,3,4].reject {|n| n%2==0}

> [1, 3]

Like select method with inverse.

January 31, 2012 - (v3.0.0 - v3.1.0)
0 thanks

undocumented events/symbol/types

@nessur is right: drop the “before_” and “after_” prefixes and you have the possible values for the uninformatively named `symbol` param: create, save, update and validate.

So to remove all validations for a model do: `reset_callbacks :validate`

January 31, 2012
2 thanks

To use with factory_girl and prevent leaking file handles

As insane-dreamer noted, to use with factory_girl:

Factory :video_file do
  file { fixture_file_upload 'test.png', 'image/png' }

However, I ran into an issue where one of our spec’s was creating a few hundred files and would then crash with:

Errno::EMFILE: Too many open files

If you look at the source code for fixture_file_upload, it creates a new file. I don’t know if these files were being leaked, or just not getting garbage collected, but they never got closed. Eventually they just piled up and the process hit the max open file limit.

So, you need to explicitly call close on the file handle returned from fixture_file_upload. However, you can’t call close in the block immediately after fixture_file_upload because the factory needs to be able to read from the stream in order to properly initialize ‘file’.

Instead what I had to do was perform the close on the ‘proxy’ in an after_create block, like so

Factory :video_file do
  file { fixture_file_upload 'test.png', 'image/png' }

  after_create do |video, proxy|
January 30, 2012
2 thanks

Requires a Block.

Just a little heads up here because it’s not obvious.

This requires a block to be passed to it.

Example Usage

say_with_time "Reverting all service rates to nil." do
  Service.update_all( :rate, nil )

# Output
-- Reverting all service rates to nil.
   -> 0.3451s
   -> 2233 rows
January 27, 2012
0 thanks

Example Usage

End of Day for Any Date

DateTime.new( 2011, 01, 01 )
# => Sat, 01 Jan 2011 00:00:00 +0000

DateTime.new( 2011, 01, 01 ).end_of_day
# => Sat, 01 Jan 2011 23:59:59 +0000

With Local Timezone

DateTime.civil_from_format( :local, 2011, 01, 01 ).end_of_day
# => Sat, 01 Jan 2011 23:59:59 -0700
January 20, 2012 - (>= v3.0.0)
0 thanks

Stubs Logger in rspec

Let we have a module like below:

module MyModule
  class << self
    def logger
      @logger ||= Logger.new(File.join(Rails.root, "log", "my_gem_#{Rails.env}.log"))

To use this logger just type:

MyModule.logger.info "This is a log line"

To stub in tests use (for rspec):

require 'active_support/log_subscriber/test_helper'

RSpec.configure do |config|
   config.include ActiveSupport::LogSubscriber::TestHelper

   config.before do

Usefull in testing when you don’t like to log anything.

January 20, 2012
1 thank


Check if id column exists in users table

ActiveRecord::Base.connection.column_exists?(:users, :id)
January 19, 2012
1 thank

Pluralize with Text.


1 person or 3 people

Use a View Helper

pluralize( 1, 'person' )
# => 1 person

pluralize( 2, 'person' )
# => 2 people

# In practice.
pluralize( Person.count, 'person' )



January 19, 2012 - (v3.0.0 - v3.1.0)
0 thanks


I came across the following situation An article has a history of friendly url being that the foreign key that represents the value of the article’s id in the table is called Friend url_id then in that case:

Article.joins(“INNER JOIN friends ON articles.id = friends.url_id”).where(“friends.url like ? ”, url)

if the column url_id was renamed for artigo_id would be easier

Article.joins(:friend).where(“friends.url like ? ”, url)

January 16, 2012
0 thanks

form_for with :as routing

The following will not work if your post model is routed with the :as option:


Instead, use the helper with your custom name:

form_for(@post, :url => edit_renamedpost_path(@post))
January 16, 2012
1 thank

Model objects routed with :as

When providing a model object, url_for will not work if the model’s routes are named using the :as option. You can instead use the named helper methods (posts_path, post_path(:id), etc.).

January 10, 2012
0 thanks

with a params constant

If you want to have a params with the same value on all of the urls in this namespace, you can write this :

with a constant param :admin set to true

namespace :admin, :admin => true do
  resources :posts

all of the urls like /admin/post have a param :admin with the value true.

It works also with :

scope 'admin', :admin => true do

match 'administration', :admin => true => 'posts#index'

get 'administration', :admin => true