govinda-fichtner / db-populate Goto Github PK

db_populate
===========
db_populate is an answer to the question "how do I get seed data into a Rails application?" 
Seed data is normally the contents of lookup tables that are essential to the normal
functioning of your application: lists of roles, administrative accounts, choices for
dropdown boxes, and so on.

The inspiration (and some of the code) for this plugin come from a blog entry by Luke
Francl (http://railspikes.com/2008/2/1/loading-seed-data) that looked at some of the
available alternatives for loading seed data. Some more of the code came from Josh
Knowles' db_populate plugin (http://code.google.com/p/db-populate/). But I didn't like 
having to assemble bits, and had some ideas to extend it, and...well, you know how it
goes.

Using db_populate
=================
The basic idea behind db_populate is simple: to put seed data in your application's
tables, it executes ruby code. The code needs to be in a specific place, and there's a 
helper to make it easier to create and update consistent seed data. Then there are a
couple of rake tasks. That's it.

Setting up for db_populate
==========================

To get started with db_populate, create the folder db/populate in your Rails application.
Any code you put in this folder will be run by db_populate. Optionally, you can create
subfolders for your Rails environments, just as you can with config files. db_populate
executes all of the top-level populate files first, followed by any environment-specific
populate files, sorting each list by name. 

Files in the 'shared' subfolder will be run for all environments, sorted with the other
files for that environment. Files in the 'after' subfolder will be run after any environment.

So, for example, with 6 files in the production environment, db_populate would order this way:

db/populate/01_roles.rb
db/populate/02_services.rb
db/populate/production/01_users.rb    
db/populate/shared/02_options.db
db/populate/production/03_accounts.rb
db/populate/after/01_cleanup.rb

Within each file, you can place whatever ruby code you like. To help create consistent
records, db_populate adds create_or_update to ActiveRecord::Base. This method looks up
a record by ID; if the record exists, it is updated, and if it doesn't, it is created. Using
this technique means that you can edit and re-run your db_populate tasks without damaging 
data that have already been loaded once. For example, assuming your roles table has already
been populated, a db_populate file to create an administrative user might look like this:

user = User.create_or_update(:id => 1, :login => "admin", :email => "[email protected]", 
 :name => "Site Administrator", :password => "admin", :password_confirmation => "admin")
role = Role.find_by_rolename('administrator')
Permission.create_or_update(:id => 1, :role_id => role.id, :user_id => user.id)

If you change your mind about the name for the site administrator, you can just edit the data
and re-run the task.

By default, create_or_update validates the data (using your model's validations) to ensure that 
the data in the database is good. You can turn off validations and attr_accessible checks 
by passing :perform_validations => false as one of the options to create_or_update:

user = User.create_or_update(:id => 1, :login => "admin", :email => "BOGUS", 
 :name => "Site Administrator", :password => "admin", :password_confirmation => "admin", :perform_validations => false)

db_populate rake tasks
======================
db_populate includes three rake tasks:

rake db:populate loads all of the data for the current environment
rake db:migrate_and_populate is the same as calling rake db:migrate followed by rake db:populate
rake db:reset_and_populate is the same as calling rake db:reset followed by rake db:populate

gem usage
=========
If you install db_populate as a gem, you need to add this line to your project Rakefile to make
the db_populate tasks available:

require 'db_populate'

History
======= 
2010-08-27 Added /shared and /after conventions
2009-11-26 Allow create_or_update to skip validations (suggestion by Josh Sharpe)
2009-11-26 Fix gem version (bug reported by Vitaly Ignatov)
2009-09-15 Add db:reset_and_populate task
2009-05-17 Fix typo in README
2009-03-21 Patch from Ahmed El-Daly to allow PKs with names other than id
2008-10-11 Initial release