NullDB is the Null Object pattern as applied to ActiveRecord database adapters. It is a database backend that translates database interactions into no-ops. Using NullDB enables you to test your model business logic - including
after_save hooks - without ever touching a real database.
Currently supported Ruby versions: MRI 2.5.x, 2.6.x, 2.7.x, 3.X.
Experimental support provided for: JRuby
Any version of ActiveRecord >= 5.2
It is tested against AR 5.2, 6.0, 6.1 and master branch.
gem install activerecord-nulldb-adapter
Once installed, NullDB can be used much like any other ActiveRecord database adapter:
ActiveRecord::Base.establish_connection :adapter => :nulldb
NullDB needs to know where you keep your schema file in order to reflect table metadata. By default it looks in RAILS_ROOT/db/schema.rb. You can override that by setting the
ActiveRecord::Base.establish_connection :adapter => :nulldb, :schema => 'foo/myschema.rb'
NullDB comes with RSpec integration. To replace the database with NullDB in all of your specs, put the following in your spec/rails_helper:
require 'nulldb_rspec' include NullDB::RSpec::NullifiedDatabase
after you load your rails environment
require File.expand_path('../config/environment', __dir__)
otherwise you will encounter issues such as (bug).
Or if you just want to use NullDB in a specific spec context, you can include the same module inside a context:
require 'nulldb_rspec' describe Employee, "with access to the database" do fixtures :employees # ... end describe Employee, "with NullDB" do include NullDB::RSpec::NullifiedDatabase # ... end
If you want to have NullDB enabled by default but disabled for particular contexts then (see this post)
NullDB::Rspec provides some custom matcher support for verifying expectations about interactions with the database:
describe Employee do include NullDB::RSpec::NullifiedDatabase it "should cause an insert statement to be executed" do Employee.create! Employee.connection.should have_executed(:insert) end end
UnitRecord-style verification that no database calls have been made at all can be achieved by using the special
describe "stuff that shouldn't touch the database" do after :each do Employee.connection.should_not have_executed(:anything) end # ... end
You can also experiment with putting NullDB in your database.yml:
unit_test: adapter: nulldb
However, due to the way Rails hard-codes specific database adapters into its standard Rake tasks, you may find that this generates unexpected and difficult-to-debug behavior. Workarounds for this are under development.
There are a number of advantages to writing unit tests that never touch the database. The biggest is probably speed of execution - unit tests must be fast for test-driven development to be practical. Another is separation of concerns: unit tests should be exercising only the business logic contained in your models, not ActiveRecord. For more on why testing-sans-database is a good idea, see: www.dcmanges.com/blog/rails-unit-record-test-without-the-database.
NullDB is one way to separate your unit tests from the database. It was inspired by the ARBS and UnitRecord libraries. It differs from them in that rather than modifying parts of ActiveRecord, it implements the same [semi-]well-documented public interface that the other standard database adapters, like MySQL and SQLServer, implement. This has enabled it to evolve to support new ActiveRecord versions relatively easily.
One concrete advantage of this null-object pattern design is that it is possible with NullDB to test
after_save hooks. With NullDB, you can call
#save and all of the usual callbacks will be called - but nothing will be saved.
It is not an in-memory database. Finds will not work. Neither will
reload, currently. Test fixtures won't work either, for obvious reasons.
It has only the most rudimentery schema/migration support. Complex migrations will probably break it.
Lots of other things probably don't work. Patches welcome!
See the LICENSE file for licensing information.