I frequently find myself needing a combination of group_by, count and sort for quick statistics. Here's a method on Enumerable that combines the three:

module Enumerable
  def count_by(&block)
    group_by(&block)
      .transform_values(&:count)
      .sort_by(&:last)
      .to_h
  end
end

Just paste that snippet into a Rails console and use #count_by now!

Usage examples

• Number of email addresses by domain:

> User.all.count_by { |user| user.email.sub /^.*@/, '' }
=> { "sina.cn"=>2, ..., "hotmail.com"=>128...

Background

• before(:all) runs the block once before all of the examples.
• before(:each) runs the block once before each of your specs.

Summary

• around(:suite) does not exist.
• around(:all) runs after before(:all) and before after(:all).
• around(:each) runs before before(:each) and after after(:each).

As this is not 100% obvious (and not yet documented) it is written down in this card. In RSpec 3 :each has the alias :example and :all the alias :context.

Example

RSpec.configure do |config|
  ...

There are many approaches out there how you can import data from a legacy application to a new application. Here is an approach which opens two database connections and uses active record for the legacy system, too:

1. Add you database information to you config/database.yml.

data_migration:
  database: your_application_data_migration

2. Create a separate application record for the data migration, e.g. in app/data_migration/migration_record.rb. You will need to create an app/data_migration.rb class first.

class DataMig...

Native promises have no methods to inspect their state.

You can use the promiseState function below to check whether a promise is fulfilled, rejected or still pending:

promiseState(promise, function(state) {
  // `state` now either "pending", "fulfilled" or "rejected"
});

Note that the callback passed to promiseState will be called asynchronously in the next [microtask](https://jakearchibald.com/2015/tasks-microtasks-queues-and-schedules/...

Use Traim to build a RESTful API for your ActiveRecord models with very little code.

Traim assumes your API resources will map 1:1 to your ActiveRecord models and database tables. This assumption usually falls apart after a few months into a project, so be ready to replace your Traim API with something more expressive afterwards.

Traim outputs a Rack application which you can either serve standalone or mount into your Rails app.

Sometimes you add Paperclip image styles, sometimes you remove some. In order to only keep the files you actually need, you should remove stale Paperclip styles from your server.

This script has been used in production successfully. Use at your own risk.

# Config #######################################################################
delete_styles = [:gallery, :thumbnail, :whatever]
scope = YourModel # A scope on the class with #has_attached_file
attachment_name = :image # First argument of #has_attached_file
noop ...

In the ruby shell (IRB) and rails console the return value of the previous command is saved in _ (underscore). This might come in handy if you forgot to save the value to a variable and further want to use it.

Example:

irb(main):001:0> 1 + 2
=> 3
irb(main):002:0> _
=> 3
irb(main):003:0> a = _
=> 3

You can not use the hash_including argument matcher with a nested hash:

describe 'user' do
  let(:user) { {id: 1, name: 'Foo', thread: {id: 1, title: 'Bar'} }

  it do 
    expect(user).to match(
      hash_including(
        id: 1, thread: {id: 1}
      )
    )
  end
end  

The example will fail and returns a not very helpful error message:

expected {:id => 1, :name => "Foo", :thread => {:id => 1, :title => "Bar"}} to...

ActiveRecord::RecordNotFound errors provide quite meaningful error messages that can provide some insight on application details. Consider the following:

ActiveRecord::RecordNotFound: Couldn't find Organisation::Membership with 'id'=12 [WHERE "organisation_memberships"."user_id" = 1]

You should probably not simply render those error messages to the user directly. Instead you you might want to re-raise your own errors. ActiveRecord::RecordNotFound provides you with methods :model and :id where you can get information about w...

This card will show you a cool way to define a class using Struct.new.
A common usecase for Structs are temporary data structures which just hold state and don't provide behaviour. In many cases you could use a simple hash as a data structure instead. However, a Struct provides you with a nice constructor, attribute accessors and complains if you try to access undefined attributes. Structs are easy to compare (by attributes). A struct gives meaning to the data.

Disclaimer

Structs are great...

Shoulda Matchers don't provide canditional validations (validations with if: option). Here is how to write tests for the condition:

Class:

class Employee < ActiveRecored::Base
  validates :office, presence: true, if: manager?
  
  def manager?
    ...
  end
end

Test:

describe Employee do

  describe '#office' do
    
    context 'is a manager' do
      before { allow(subject).to receive(:manager?).and_return(true) }
      it { is_expected.to validate_presence_o...

When using GROUP BY, MySQL now complains if the SELECT includes columns which are not part of the GROUP BY.

Reason:

There could be multiple values for those columns per group but only one value can be picked for the results.

The default behaviour of MySQL prior to version 5.7 will not complain and arbitrarily choose a value. But this leads to non-deterministic results. So MySQL now has enabled the only_full_group_by setting by default to prevent this.

In Rails this could lead to some trouble, because scopes do not have sp...

Nowadays it is fairly easy to intercept and modify mails globally before they are sent. All you have to do is register an interceptor class which responds to .delivering_email(message). This card will show you two common use cases.

Subject prefix:

Usually you want to prefix the subject line of emails with the current environment (except production) so you can differentiate between production mails and mails from other environments. Of course a...

Rails supports time zones, but there are several pitfalls. Most importantly because Time.now and Time.current are completely different things and code from gems might use one or the other.

Especially configuring an application that cares only about one time zone is a bit tricky.

The following was tested on Rails 5.1 but should apply to Rails 4.2 as well.

Using only local time

Your life will be easier if your application does not need to support time zones. Disable them like this:

config.time_zone = 'Berlin' # Your local ...

Webmocks hash_including is similar to RSpec::Mocks::ArgumentMatchers#hash_including. Be aware that hash_including (webmock v3.0.1) doesn't parse integer values to String.

Without hash including you would say:

uri = URI('http://example.com/?foo=1&bar=2')
stub_request(:get, 'example.com').with(query: {foo: 1, bar: 2})
Net::HTTP.get(uri) # ===> Success

If you only want to check if foo is present you can use hash_including:

uri = URI('http://example.com/?foo=1&bar=2')
stub_request(:get, 'example.com').with(query: hash_i...

TLDR: Browser implementations of HTTP/2 push are horrible. You might end up with worse performance than without pushing. However, the article includes a great explanation of how HTTP/2 push are supposed to integrate with browser APIs.

Dump this method into your Ruby console to quickly print data in columns. This is helpful for e.g. comparing attributes of a set of Rails records.

def tp(objects, *method_names)
  terminal_width = `tput cols`.to_i
  cols = objects.count + 1 # Label column
  col_width = (terminal_width / cols) - 1 # Column spacing

  Array(method_names).map do |method_name|
    cells = objects.map{ |o| o.send(method_name).inspect }
    cells.unshift(method_name)

    puts cells.map{ |cell| cell.to_s.ljust(col_width) }.join ' '
  end

  nil
end

Usag...

Database connections are not thread-safe. That's why ActiveRecord uses a separate database connection for each thread.

For instance, the following code uses 3 database connections:

3.times do
  Thread.new do
    User.first # first database access makes a new connection
  end
end

These three connections will remain connected to the database server after the threads terminate. This only affects threads that use ActiveRecord.

You can rely on Rails' various clean-up mechanisms to release connections, as outlined below. This may...

This card compares patterns to store trees in a relation database like MySQL or PostgreSQL. Implementation examples are for the ActiveRecord ORM used with Ruby on Rails, but the techniques can be implemented in any language or framework.

We will be using this example tree (from the acts_as_nested_set docs):

root
|
+-- Child 1
|   |
|   +-- Child 1.1
|   |
|   +-- Child 1.2
|
+-- ...

The parallel-gem is quite easy to use and can speed up rendering time if you want to render the same partial multiple times (e.g. for rendering long lists of things).
If your parallelized code talks to the database, you should ensure not to leak database connections.

Consider you want to render a list of groups with their members as json. You can use a partial for the rendering of group members, b...

When you load a dump for development, records may reference images that are not available on your machine.

Requests to those images may end up on your application, e.g. if a catch-all route is defined that leads to a controller doing some heavy lifting. On pages with lots of missing images, this slows down development response times.

You can fix that by defining a Rails route like this:

if Rails.env.development?
  scope format: true, constraints: { format: /jpg|png|gif/ } do
    get '/*anything', to: proc { [404, {}, ['']] }

...