Seed Migrations allow you treat seed data like migration data:
- each seed is imported only once
- seeds are imported in a specific order
A rake db:seed command is provided for pre-2.3.4 Rails users.
db-populate and seed-fu seemed overkill (though create_or_update is neat). Seed-fu didn't appear to support ordering the imports, which makes handling FK relations complicated (i.e. enabling and disabling all constraints.)
While I was flip flopping like Al Gore, I found myself in a situation where the code had to be custom. I saw a post by Mike Gunderloy on A Fresh Cup referring to seed data support in Rails 3. A quick look revealed it to not be much of a feature, but a basic rake task to run a seeds.rb file which just happens to contain seed data (or not.)
The only vaguely cool feature is a migration-style record of applied seed data imports. This allows you to run the db:seed as many times as you like without repeating inserts.
script/plugin git://github.com/n3bulous/seed_migrations.git- Copy db/seeds.rb to your Rails' db directory.
Obviously, if you already have a db/seeds.rb you'll have to migrate. The most simple way to achieve this is:
script/generate seed original_seeds- copy the current seeds.rb file into it
- prepend the seed file name (sans the .rb extension) to db/seeds/seed_order.yaml
- if the existing seeds can't be re-run, insert the seed filename into the seedings table before running
rake db:seedfor the first time.
./script/generate seed <a-descriptive-name>- Add Ruby and Rails code to the generated file in the db/seeds directory, e.g.:
cities = City.create([{ :name => 'Chicago' }, { :name => 'Copenhagen' }])
Mayor.create(:name => 'Daley', :city => cities.first)
rake db:seed
The entries in seed_order.yaml are used as the key in the seedings table that keeps track of which ones have been inserted.
- db:seed:reset/up/down but this might be overkill.
- automate migrating from an pre-existing seeds.rb file.
- Patrick Reagan
- improve compatibility of the generator code
Caveat Emptor