Rake unconsistent behaviour between dev and test? - ruby-on-rails

Question

I develop a RoR application using a PostgreSQL database, based on this database.yml definition:
# PostGre databases
default: &default
host : localhost
adapter: postgresql
encoding: unicode
pool: 5
username: keyman
password: keymanApp
schema_search_path: "keyman"
development:
<<: *default
database: keyman_dev
test:
<<: *default
database: keyman_test
I created a small Rake routine, so I can easily drop and create my postgreSQL database, including the schema I work with:
namespace :db do
desc 'Create database schemas before going for the first migration'
task init: ['db:drop','db:create'] do
ActiveRecord::Base.connection.execute("CREATE SCHEMA keyman AUTHORIZATION keyman")
puts 'Database initialised'
end
end
When I run rake db:init, it is executed both on dev and test environments:
$ rake db:init
Dropped database 'keyman_dev'
Dropped database 'keyman_test'
Created database 'keyman_dev'
Created database 'keyman_test'
Database initialised
But the result is not the same: the schema 'keyman' is created for the keyman_dev database, but not for the keyman_test database.
I need to run explicitly rake db:init RAILS_ENV=test to get the schema created on the test database.
It sounds strange to me! Do you have an explanation?
Thanks

Answer 1

when running bin/rake -T db we can see the following descriptions for db:create and db:drop
rake db:create # Creates the database from DATABASE_URL or config/database.yml for the current RAILS_ENV (use db:create:all to create all databases in the config). Without RAILS_ENV or when RAILS_ENV is development, it defaults to creating the development and test databases
rake db:drop # Drops the database from DATABASE_URL or config/database.yml for the current RAILS_ENV (use db:drop:all to drop all databases in the config). Without RAILS_ENV or when RAILS_ENV is development, it defaults to dropping the development and test databases
So, when you run those tasks in the development environment, they both do the development and test databases, but that doesn't mean that other tasks (such as your custom task) automatically run in both environments, you're task is still only running in the development environment.
Something like this in your rake task should get that query to run in both tables, though this is untested.
environments = Rails.env.development? ? [:development, :test] : [Rails.env.to_sym]
environments.each do |env|
ActiveRecord::Base.establish_connection(env)
# do something
end