Skip to content


Subversion checkout URL

You can clone with
Download ZIP
A little random, unique token generator for Mongoid documents.

Merge pull request #62 from thetron/patch/8

Added support for tokens replacing default the `_id` field
latest commit 50edba6d40
@thetron authored

Mongoid::Token - Short snappy tokens for Mongoid documents

Build Status Code Climate

This library is a quick and simple way to generate unique, random tokens for your mongoid documents, in the cases where you can't, or don't want to use slugs, or the default MongoDB ObjectIDs.

Mongoid::Token can help turn this:

Into something more like this:

Getting started

In your gemfile, add:

gem 'mongoid_token', '~> 2.0.0'

Then update your bundle

$ bundle install

In your Mongoid documents, just add include Mongoid::Token and the token method will take care of all the setup, like so:

class Person
  include Mongoid::Document
  include Mongoid::Token

  field :name


And that's it! There's lots of configuration options too - which are all listed below. By default, the token method will create tokens 4 characters long, containing random alphanumeric characters.

Note: Mongoid::Token leverages Mongoid's 'safe mode' by automatically creating a unique index on your documents using the token field. In order to take advantage of this feature (and ensure that your documents always have unique tokens) remember to create your indexes.

Using without Rails

If you're using Mongoid without Rails, remember to:

require 'mongoid_token'

in your app


By default, the gem will add a convenience find_by_[token name] method, to make it a tiny bit simpler to query by your tokens. If you'd rather it didn't, you can disable these with the skip_finders configuration option.

The methods accept either a single token, or an array of tokens:

Video.find_by_token("x3v98") # If your token was named 'token'
Account.find_by_account_number(["ACC-123456", "ACC-567890"]) # If your token was named 'account_number'



As of Mongoid::Token 2.0.0, you can now choose between two different systems for managing how your tokens look.

For simple setup, you can use combination of the length and contains, which modify the length and types of characters to use.

For when you need to generate more complex tokens, you can use the pattern option, which allows for very low-level control of the precise structure of your tokens, as well as allowing for static strings, like prefixes, infixes or suffixes.

Length (:length)

This one is easy, it's just an integer.


token :length => 6 # Tokens are now of length 6
token :length => 12 # Whow, whow, whow. Slow down egghead.

You get the idea.

The only caveat here is that if used in combination with the :contains => :numeric option, tokens may vary in length up to the specified length.

Contains (:contains)

Contains has 7 different options:

  • :alphanumeric - contains uppercase & lowercase characters, as well as numbers
  • :alpha - contains only uppercase & lowercase characters
  • :alpha_upper - contains only uppercase letters
  • :alpha_lower - contains only lowercase letters
  • :numeric - integer, length may be shorter than :length
  • :fixed_numeric - integer, but will always be of length :length
  • :fixed_numeric_no_leading_zeros - same as :fixed_numeric, but will never start with zeros
  • :fixed_hex_numeric - hex integer, but will always be of length :length
  • :fixed_hex_numeric_no_leading_zeros - same as :fixed_hex_numeric, but will never start with zeros


token :contains => :alpha_upper, :length => 8
token :contains => :fixed_numeric

Patterns (:pattern)

New in 2.0.0, patterns allow you fine-grained control over how your tokens look. It's great for generating random data that has a requirements to also have some basic structure. If you use the :pattern option, it will override both the :length and :contains options.

This was designed to operate in a similar way to something like strftime, if the syntax offends you - please open an issue, I'd love to get some feedback here and better refine how these are generated.

Any characters in the string are treated as static, except those that are proceeded by a %. Those special characters represent a single, randomly generated character, and are as follows:

  • %s - any uppercase, lowercase, or numeric character
  • %w - any uppercase, or lowercase character
  • %c - any lowercase character
  • %C - any uppercase character
  • %d - any digit
  • %D - any non-zero digit


token :pattern => "PRE-%C%C-%d%d%d%d" # Generates something like: 'PRE-ND-3485'

You can also add a repetition modifier, which can help improve readability on more complex patterns. You simply add any integer after the letter.


token :pattern => "APP-%d6" # Generates something like; "APP-638924"

Field Name (:field_name)

This allows you to change the field name used by Mongoid::Token (default is :token). This is particularly handy when you're wanting to use multiple tokens one a single document.


token :length => 6
token :field_name => :sharing_token, :length => 12
token :field_name => :yet_another

Skip Finders (:skip_finders)

This will prevent the gem from creating the extra find_by_* methods.


token :skip_finders => true

Override to_param (:override_to_param)

By default, Mongoid::Token will override to_param, to make it an easy drop-in replacement for the default ObjectIDs. If needed, you can turn this behaviour off:


token :override_to_param => false

Retry Count (:retry_count)

In the event of a token collision, this gem will attempt to try three more times before raising a Mongoid::Token::CollisionRetriesExceeded error. If you're wanting it to try harder, or less hard, then this option is for you.


token :retry_count => 9
token :retry_count => 0


If you find a problem, please submit an issue (and a failing test, if you can). Pull requests and feature requests are always welcome and greatly appreciated.

Thanks to everyone that has contributed to this gem over the past year. Many, many thanks - you guys rawk.


Thanks to everyone who has provided support for this gem over the years. In particular: olliem, msolli, siong1987, stephan778, eagleas, and jamesotron.

Something went wrong with that request. Please try again.