Skip to content
Persisted data with yaml
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.



YAML Record is a data persistence library that complies with the ActiveModel API. Using YAMLRecord should be familiar to anyone that has used ActiveRecord before to manage your database. Using this library, the data is persisted in a YAML backed file.


Why a YAML-based persistence store? In certain situations, there are collections of simple data in which there are very few records which are by nature infrequently accessed and that are ideally able to be scanned easily within a text file. These can include a simple contact form, landing page interest, feedback forms, surveys, team pages, etc where there is simply no need for the overhead of a fully persisted database solution.

There are many cases where YAMLRecord is not the correct persistence strategy. Any collection that is going to have substantial number of records, will be frequently updated, or is accessible by a large volume of users should not be stored in a YAML text file for obvious reasons. However, for specific cases, the convenience of storing things in a simple text file becomes apparent. Being able to access the text file data as if the records were in a familiar database ORM has many conveniences and advantages such as keeping the controllers standard and leveraging existing ORM knowledge.


Install using rubygems:

gem install yaml_record

Or add gem to your Gemfile:

# Gemfile
gem 'yaml_record'
# OR if you're using Rails 3.1
gem 'yaml_record', :git => "", :branch => "rails31"



Create any ruby object and inherit from YamlRecord:Base to define a type:

class Post < YamlRecord::Base
  # Declare your properties
  properties :title, :body, :user_id

  # Declare your adapter (local by default)
  adapter :local # or :redis
  # Declare source file path
  source Rails.root.join("config/posts")

Use this new object the same way as any ActiveRecord object.


Retrieve the collection:

Post.all => [@p1, @p2]

Retrieve item by id:

Post.find("a1b2") => @p1

Retrieve by attribute:

Post.find_by_attribute(:title, "some title") => @p


Initialize post:

@p = => "...", :body => "...", :user_id => 5)

Create post:

@p = Post.create(:title => "...", :body => "...", :user_id => 6)


Update attributes using the expected method:

@p.update_attributes(:title => "new title")


Destroy a given record:



Access attributes:

@p = Post.find("a1b2")
@p.title => "..."

Assign attributes:

@p.title = "new title"


Create callbacks:

class Submission < YamlRecord::Base
  # ...
  before_create :do_something # or before_save, before_destroy, ...

  def do_something
    # something here

Storage Adapters

YAMLRecord supports pluggable storage adapters that control the storage engine used for the YAML data. By default, the adapter used is the local store which writes a file (specified by source path) to the local system. There are currently two available adapters: Local and Redis.

To configure the adapter, you can simply declare within the object:

class Submission < YamlRecord::Base
  adapter :redis, $redis # Second parameter is the redis client instance
  source "contacts" # stores yaml namespaced as 'yaml_record:contacts' in redis

Each storage adapter only defines a read and write interface and is easy to create. Checkout the redis adapter for an example of how simple they are to define. Feel free to create additional adapters and send them to us via a pull request.


Imagine a simple contact form that accepts a name and email from a user along with a body:

class Submission < YamlRecord::Base
  # Declare your properties
  properties :name, :email, :body

  # Declare your adapter (local by default)
  adapter :local # or :redis
  # Declare source file path (config/contact.yml)
  source Rails.root.join("config/contact")

Once we define the Contact model, we can setup a controller and form just the same as in ActiveRecord:

class SubmissionsController < AC::Base
  def create
    @submission = Submission.create(params[:submission])

  def index
    @submissions = Submission.all

  def show
    @submission = Submission.find(params[:id])

  def update
    @submission = Submission.find(params[:id])

  def destroy
    @submission = Submission.find(params[:id])

As you can see the controller appears the same as any ActiveRecord controller would and this makes managing the YAML data easy and convenient. You can even define callbacks in your object as you would in ActiveRecord:

class Submission < YamlRecord::Base
  # ...
  before_create :do_something # or before_save, before_destroy, ...

  def do_something
    # something here

And that's all! Each record will be persisted to the source file for easy access.


  • Validations should be supported validates_presence_of :name
  • Property type declarations should be available property :age, Integer


Created at Miso by Nico Taing and Nathan Esquenazi

Special thanks to Vaudoc

Contributors and patches are welcome! Please send a pull request!


There is already an excellent project for YAML persistence if you are using Datamapper. In the situation in which we were using DM and Padrino, this would surely be a better choice. But if you are using ActiveRecord and Rails, this library is a lightweight and standalone solution.


YAML Record is Copyright © 2011 Nico Taing, Miso. It is free software, and may be redistributed under the terms specified in the MIT-LICENSE file.

You can’t perform that action at this time.