FlagShihTzu

Bit fields for ActiveRecord

An extension for ActiveRecord to store a collection of boolean attributes in a single integer column as a bit field.

github.com/xing/flag_shih_tzu

This gem lets you use a single integer column in an ActiveRecord model to store a collection of boolean attributes (flags). Each flag can be used almost in the same way you would use any boolean attribute on an ActiveRecord object.

The benefits:

  • No migrations needed for new boolean attributes. This helps a lot if you have very large db-tables, on which you want to avoid ALTER TABLE whenever possible.

  • Only the one integer column needs to be indexed.

Using FlagShihTzu, you can add new boolean attributes whenever you want, without needing any migration. Just add a new flag to the has_flags call.

And just in case you are wondering what a “Shih Tzu” is: en.wikipedia.org/wiki/Shih_Tzu

Build status

Prerequisites

The gem is actively being tested with:

  • ActiveRecord versions 2.3.x, 3.0.x, 3.1.x, 3.2.x

  • MySQL, PostgreSQL and SQLite3 databases

  • Ruby 1.8.7, 1.9.2 and 1.9.3

Installation

Rails 2.x

In environment.rb:

config.gem 'flag_shih_tzu'

Then:

$ rake gems:install # use sudo if necessary

Rails 3

In Gemfile:

gem 'flag_shih_tzu'

Then:

$ bundle install

Usage

FlagShihTzu assumes that your ActiveRecord model already has an integer field to store the flags, which should be defined to not allow NULL values and should have a default value of 0 (which means all flags are initially set to false).

Defining the flags

class Spaceship < ActiveRecord::Base
  include FlagShihTzu

  has_flags 1 => :warpdrive,
            2 => :shields,
            3 => :electrolytes
end

has_flags takes a hash. The keys must be positive integers and represent the position of the bit being used to enable or disable the flag. The keys must not be changed once in use, or you will get wrong results. That is why the plugin forces you to set them explicitly. The values are symbols for the flags being created.

How it stores the values

As said, FlagShihTzu uses a single integer column to store the values for all the defined flags as a bit field.

The bit position of a flag corresponds to the given key.

This way, we can use bit operators on the stored integer value to set, unset and check individual flags.

              +---+---+---+                +---+---+---+
              |   |   |   |                |   |   |   |
Bit position  | 3 | 2 | 1 |                | 3 | 2 | 1 |
(flag key)    |   |   |   |                |   |   |   |
              +---+---+---+                +---+---+---+
              |   |   |   |                |   |   |   |
Bit value     | 4 | 2 | 1 |                | 4 | 2 | 1 |
              |   |   |   |                |   |   |   |
              +---+---+---+                +---+---+---+
              | e | s | w |                | e | s | w |
              | l | h | a |                | l | h | a |
              | e | i | r |                | e | i | r |
              | c | e | p |                | c | e | p |
              | t | l | d |                | t | l | d |
              | r | d | r |                | r | d | r |
              | o | s | i |                | o | s | i |
              | l |   | v |                | l |   | v |
              | y |   | e |                | y |   | e |
              | t |   |   |                | t |   |   |
              | e |   |   |                | e |   |   |
              | s |   |   |                | s |   |   |
              +---+---+---+                +---+---+---+
              | 1 | 1 | 0 | = 4 + 2 = 6    | 1 | 0 | 1 | = 4 + 1 = 5
              +---+---+---+                +---+---+---+

Read more about bit fields here: en.wikipedia.org/wiki/Bit_field

Using a custom column name

The default column name to store the flags is 'flags', but you can provide a custom column name using the :column option. This allows you to use different columns for separate flags:

has_flags 1 => :warpdrive,
          2 => :shields,
          3 => :electrolytes,
          :column => 'features'

has_flags 1 => :spock,
          2 => :scott,
          3 => :kirk,
          :column => 'crew'

Generated instance methods

Calling has_flags as shown above creates the following instance methods on Spaceship:

Spaceship#warpdrive
Spaceship#warpdrive?
Spaceship#warpdrive=
Spaceship#warpdrive_changed?

Spaceship#shields
Spaceship#shields?
Spaceship#shields=
Spaceship#shields_changed?

Spaceship#electrolytes
Spaceship#electrolytes?
Spaceship#electrolytes=
Spaceship#electrolytes_changed?

Opionally, you can set the :bang_methods option to true to enable the bang methods:

Spaceship#electrolytes!
Spaceship#not_electrolytes!

which respectively enables or disables the electrolytes flag.

Generated named scopes

The following named scopes become available:

Spaceship.warpdrive         # :conditions => "(spaceships.flags in (1,3,5,7))"
Spaceship.not_warpdrive     # :conditions => "(spaceships.flags not in (1,3,5,7))"
Spaceship.shields           # :conditions => "(spaceships.flags in (2,3,6,7))"
Spaceship.not_shields       # :conditions => "(spaceships.flags not in (2,3,6,7))"
Spaceship.electrolytes      # :conditions => "(spaceships.flags in (4,5,6,7))"
Spaceship.not_electrolytes  # :conditions => "(spaceships.flags not in (4,5,6,7))"

If you do not want the named scopes to be defined, set the :named_scopes option to false when calling has_flags:

has_flags 1 => :warpdrive, 2 => :shields, 3 => :electrolytes, :named_scopes => false

In a Rails 3 application, FlagShihTzu will use scope internally to generate the scopes. The option on has_flags is still named :named_scopes however.

Examples for using the generated methods

enterprise = Spaceship.new
enterprise.warpdrive = true
enterprise.shields = true
enterprise.electrolytes = false
enterprise.save

if enterprise.shields?
  ...
end

Spaceship.warpdrive.find(:all)
Spaceship.not_electrolytes.count
...

Support for manually building conditions

The following class methods may support you when manually building ActiveRecord conditions:

Spaceship.warpdrive_condition         # "(spaceships.flags in (1,3,5,7))"
Spaceship.not_warpdrive_condition     # "(spaceships.flags not in (1,3,5,7))"
Spaceship.shields_condition           # "(spaceships.flags in (2,3,6,7))"
Spaceship.not_shields_condition       # "(spaceships.flags not in (2,3,6,7))"
Spaceship.electrolytes_condition      # "(spaceships.flags in (4,5,6,7))"
Spaceship.not_electrolytes_condition  # "(spaceships.flags not in (4,5,6,7))"

These methods also accept a :table_alias option that can be used when generating SQL that references the same table more than once:

Spaceship.shields_condition(:table_alias => 'evil_spaceships') # "(evil_spaceships.flags in (2,3,6,7))"

Choosing a query mode

While the default way of building the SQL conditions uses an IN() list (as shown above), this approach will not work well for a high number of flags, as the value list for IN() grows.

For MySQL, depending on your MySQL settings, this can even hit the 'max_allowed_packet' limit with the generated query.

In this case, consider changing the flag query mode to :bit_operator instead of :in_list, like so:

has_flags 1 => :warpdrive,
          2 => :shields,
          :flag_query_mode => :bit_operator

This will modify the generated condition and named_scope methods to use bit operators in the SQL instead of an IN() list:

Spaceship.warpdrive_condition     # "(spaceships.flags & 1 = 1)",
Spaceship.not_warpdrive_condition # "(spaceships.flags & 1 = 0)",
Spaceship.shields_condition       # "(spaceships.flags & 2 = 2)",
Spaceship.not_shields_condition   # "(spaceships.flags & 2 = 0)",

Spaceship.warpdrive     # :conditions => "(spaceships.flags & 1 = 1)"
Spaceship.not_warpdrive # :conditions => "(spaceships.flags & 1 = 0)"
Spaceship.shields       # :conditions => "(spaceships.flags & 2 = 2)"
Spaceship.not_shields   # :conditions => "(spaceships.flags & 2 = 0)"

The drawback is that due to the bit operator, this query can not use an index on the flags column.

Running the gem tests

First, make sure all required gems are installed:

$ bundle install

The default rake test task will run the tests against the currently locked ActiveRecord version (see Gemfile.lock):

$ bundle exec rake test

If you want to run the tests against all supported ActiveRecord versions:

$ bundle exec rake test:all

This will internally use bundler to load specific ActiveRecord versions before executing the tests (see gemfiles/), e.g.:

$ BUNDLE_GEMFILE='gemfiles/Gemfile.activerecord-3.1.x' bundle exec rake test

All tests will use an in-memory sqlite database by default. If you want to use a different database, see test/database.yml, install the required adapter gem and use the DB environment variable to specify which config from test/database.yml to use, e.g.:

$ DB=mysql bundle exec rake

Authors

Patryk Peszko, Sebastian Roebke, David Anderson, Tim Payton and a helpful group of contributors. Thanks!

Please find out more about our work in our Devblog.

License

The MIT License

Copyright © 2011 XING AG

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.