I've followed the raging debate over at Auto-load the seed data from db/seeds.rb with rake. Die-hards maintain that you should never load seed data for tests, but I take the more moderate stance that there are occasions where you might want to load seed data for specific tests, e.g. verifying that the seed data exists.
Unlike some answers given here, I do not recommend unconditionally loading the seeds from inside your spec_helper file. Instead, you can load your seeds using before :each
or before :all
inside just those test files that need the seeds, e.g.:
describe "db seed tests" do
before(:each) do
load "#{Rails.root}/db/seeds.rb"
end
...your test code here...
end
update
As @marzapower points out, if you go this route, your seeds.db file should clear each table before creating entries or use find_or_create_by
methods. (Hint: the former is faster and more reliable.) This will prevent duplicate entries if you load the seeds.db file more than once.