With the Spring Framework you can configure which Beans (components) are loaded into the Application Context in different ways.

This is particularly useful for testing, loading all components of your application may not be necessary for certain types of test.

This article is about the case when you want to test your data access layer and at the same time:

  • keep your tests independent from each other
  • load a minimal set of components to the application context

Test dependencies

There are legitimate reasons to write test that depend on other tests. This article is about the case where you want test independence. Some of the associated problems of dependent tests are:

  • running the test in different environments leads to different results
  • unpredictable test results when multiple tests reuse the same data source

Using auto-configured JPA tests, enter @DataJpaTest

Spring provides the annotation @DataJpaTest, it will create a “slice” which will ensure the test class only loads components relevant to the execution of test involving JPA components and it’s dependencies.

Here’s a class diagram for the @DataJpaTest annotation:

Class diagram for @DataJpaTest

The annotation packs lots of features:

  • automated transaction management, provided by @Transactional, transactions roll back at the end of each test
  • loads only JPA related components
  • an auto-configured in-memory database ready for testing
  • an EntityManager ready for testing, it allows you to persist entities without having to use the @Repository interface

Using the annotation is quite simple, here’s a small example, adapted from the Spring documentation:

    @RunWith(SpringRunner.class)
    @DataJpaTest
    public class ExampleTransactionalTests {
        @Autowired
        private TestEntityManager entityManager;

        @Autowired
        private UserRepository repository;

        @Test
        public void testExample() throws Exception {
            this.entityManager.persist(new User("sboot", "1234"));
            User user = this.repository.findByUsername("sboot");
            assertThat(user.getUsername()).isEqualTo("sboot");
            assertThat(user.getVin()).isEqualTo("1234");
        }
    }

Alternatives to auto-configured tests

Your own set-up and tear-down

You can always write your own set-up and tear-down methods to ensure every tests uses a data source in a pristine state, however it comes with trade-offs, namely:

  • it’s code that you need to write and maintain
  • test execution can be optimized to be faster

Destroy and re-create the application context

Spring also provides the annotation @DirtiesContext as a means to mark the application context as “dirty”, Spring will recreate the context according to your configuration – before/after each test/test class –. It also comes with trade-offs, namely:

  • re-creating the application context is expensive and your test will take longer
  • there is no code that you have to maintain

References

Google search terms

  • @datajpatest site:docs.spring.io
  • test slice jpa site:docs.spring.io