Spring DBUnit provides integration between the Spring testing framework and the popular DBUnit project. It allows to set up and tear down database tables using simple annotations as well as checking expected table contents once a test completes.
The project can be configured to run DBUnit tests using a Spring TestExecutionListener.
This project was forked from springtestdbunit/spring-test-dbunit in order to continue its maintenance.
The versioning of this project follows the one of the Spring Framework:
- Major version matches the one of the Spring Framework
- Minor version is used for dependency upgrades
- Patch version is used for bug fixing
The dependencies used are:
- Spring Framework 6
- DBUnit 2.7
- JUnit 5
All artefacts of this project are available on Maven's central repository, which makes it easy to use in your projects.
If you are using Maven, simply declare the following dependencies:
-
spring-test-dbunit-core:
<dependency>
<groupId>com.github.ppodgorsek</groupId>
<artifactId>spring-test-dbunit-core</artifactId>
<version>${spring-test-dbunit.version}</version>
</dependency> -
spring-test-dbunit-sample:
<dependency>
<groupId>com.github.ppodgorsek</groupId>
<artifactId>spring-test-dbunit-sample</artifactId>
<version>${spring-test-dbunit.version}</version>
</dependency>
spring-test-dbunit-core is the only mandatory one, the others are optional.
To have Spring process DBUnit annotations you must first configure your tests to use the DbUnitTestExecutionListener class. To do this you need to use the Spring @TestExecutionListeners annotation. Generally, as well as DbUnitTestExecutionListener, you will also want to include the standard Spring listeners as well. Here are the annotations for a typical JUnit 5 test:
@SpringJUnitConfig
@ContextConfiguration
@TestExecutionListeners({ DependencyInjectionTestExecutionListener.class,
DirtiesContextTestExecutionListener.class,
TransactionalTestExecutionListener.class,
DbUnitTestExecutionListener.class })
See the Spring JavaDocs for details of the standard listeners.
In order to access the database, Spring DBUnit requires a bean to be registered in your test context. By default a bean named dbUnitDatabaseConnection or dataSource can be used (see the Advanced Configuration section below if you need to use another name). The bean can reference either a IDatabaseConnection or more typically a standard Java DataSource. Here is a typical XML configuration for accessing an in-memory Hypersonic database:
<bean id="dataSource" class="org.springframework.jdbc.datasource.DriverManagerDataSource">
<property name="driverClassName" value="org.hsqldb.jdbcDriver" />
<property name="url" value="jdbc:hsqldb:mem:paging" />
<property name="username" value="sa" />
<property name="password" value="" />
</bean>
Once you have configured the DbUnitTestExecutionListener and provided the bean to access your database you can use the DBUnit annotations.
Note: You need to complete the steps from the configuration section above before any annotations can be used. Without appropriate configuration DBUnit annotations will be silently ignored.
The @DatabaseSetup and @DatabaseTearDown annotations can be used to configure database tables before tests execute and reset them once tests have completed.
The @DatabaseSetup annotation indicates how database tables should be setup before test methods are run. The annotation can be applied to individual test methods or to a whole class. When applied at the class level the setup occurs before each method in the test. The annotation value references a file that contains the table's dataset used when resetting the database. Typically this is a standard DBUnit XML file, although it is possible to load custom formats (see below).
Here is a typical setup annotation. In this case a file named sampleData.xml is contained in the same package as the test class.
@DatabaseSetup("sampleData.xml")
It is also possible to reference specific resource locations, for example:
@DatabaseSetup("/META-INF/dbtest/sampleData.xml")
By default, the setup will perform a CLEAN_INSERT operation, this means that all data from tables referenced in the DataSet will be removed before inserting new rows. The standard DBUnit operations are supported using type attribute. See the JavaDocs for full details.
The @DatabaseTearDown annotation can be used to reset database tables once a test has completed. As with @DatabaseSetup, the annotation can be applied at the method or class level. When using @DatabaseTearDown use the value and type attributes in the same way as @DatabaseSetup.
Note: If you are running a teardown in conjunction with a @Transactional test you may need to use an alternative configuration. See the section on below.
The @ExpectedDatabase annotation can be used to verify the contents of the database once a test has completed. You would typically use this annotation when a test performs an insert, update or delete. You can apply the annotation on a single test method or a class. When applied at the class level, verification occurs after each test method.
The @ExpectedDatabase annotation takes a value attribute that references the DataSet file used to verify results. Here is a typical example:
@ExpectedDatabase("expectedData.xml")
The @ExpectedDatabase annotation supports two different modes. DatabaseAssertionMode.DEFAULT operates as any standard DbUnit test, performing a complete compare of the expected and actual datasets. DatabaseAssertionMode.NON_STRICT will ignore tables and column names which are not specified in the expected dataset but exist in the actual datasets. This can be useful during integration tests performed on live databases containing multiple tables that have many columns, so one must not specify all of them, but only the 'interesting' ones.
Note: If you are using this annotation in conjunction with a @Transactional test you may need to use an alternative configuration. See the section on below.
If you have configured DBUnit tests to run using the DbUnitTestExecutionListener and are also using the TransactionalTestExecutionListener, you may experience problems with transactions not being started before your data is setup, or being rolled back before expected results can be verified. In order to support @Transactional tests with DBUnit, you should use the TransactionDbUnitTestExecutionListener class.
Here are the annotations for a typical JUnit 5 test:
@SpringJUnitConfig
@ContextConfiguration
@Transactional
@TestExecutionListeners({ DependencyInjectionTestExecutionListener.class,
DirtiesContextTestExecutionListener.class,
TransactionDbUnitTestExecutionListener.class })
Transactions start before @DatabaseSetup and end after @DatabaseTearDown and @ExpectedDatabase.
The @DbUnitConfiguration annotation can be used if you need to configure advanced options for DBUnit.
The databaseConnection attribute allows you to specify a specific bean name from the Spring Context that contains the database connection. When not specified the names or can be used. The bean must be either an IDatabaseConnection or a DataSource.
The dataSetLoader or dataSetLoaderBean attribute allows you to specify a custom loader that will be used when reading datasets (see below). If no specific loader is specified a dbUnitDataSetLoader bean will be used from the ApplicationContext (or if no such bean exists, the FlatXmlDataSetLoader will be used).
The databaseOperationLookup attribute allows you to specify a custom lookup strategy for DBUnit database operations (see below).
It is possible to configure Spring Test DBUnit to work with multiple connections within the same test. First declare multiple DataSource or IDatabaseConnection beans in your application context. For example, here is XML configuration for two in-memory databases:
<bean id="dataSource" class="org.springframework.jdbc.datasource.DriverManagerDataSource">
<property name="driverClassName" value="org.hsqldb.jdbcDriver" />
<property name="url" value="jdbc:hsqldb:mem:paging" />
<property name="username" value="sa" />
<property name="password" value="" />
</bean>
<bean id="customerDataSource" class="org.springframework.jdbc.datasource.DriverManagerDataSource">
<property name="driverClassName" value="org.hsqldb.jdbcDriver" />
<property name="url" value="jdbc:hsqldb:mem:customers" />
<property name="username" value="sa" />
<property name="password" value="" />
</bean>
You then need to use the @DbUnitConfiguration on your test to link to the connections:
@DbUnitConfiguration(databaseConnection={"dataSource", "customerDataSource"})
The @DatabaseSetup, @DatabaseTearDown and @ExpectedDatabase annotations all have a connection attribute which can be used if you need to target a specific connection. If you don't specify a connection the first databaseConnection from @DbUnitConfiguration will be used ("dataSource" in the example above).
Spring Test DBUnit Annotations are repeatable so if you are using Java 8+ you can use several with the same test. For example:
@Test
@DatabaseSetup(value = "insert.xml")
@DatabaseSetup(connection="customerDataSource", value="insert-custs.xml")
public void testInsert() throws Exception {
// Inserts "insert.xml" into dataSource and "insert-custs.xml" into customerDataSource
// ...
}
If you are using an earlier version of Java, you will need to use one of the wrapper annotations:
@Test
@DatabaseSetups({
@DatabaseSetup(value = "insert.xml")
@DatabaseSetup(connection="customerDataSource", value="insert-custs.xml")
})
public void testInsert() throws Exception {
// Inserts "insert.xml" into dataSource and "insert-custs.xml" into customerDataSource
// ...
}
In some situations you may need to create an IDatabaseConnection with a specific DBUnit configuration. Unfortunately, the standard DBUnit DatabaseConfig class cannot be set up easily using Spring. In order to overcome this limitation, the DatabaseConfigBean provides an alternative method to configure a connection; with standard getter/setter access provided for all configuration options. The DatabaseDataSourceConnectionFactoryBean accepts a configuration property and should be used to construct the final connection. Here is a typical example:
<bean id="dbUnitDatabaseConfig" class="com.github.springtestdbunit.bean.DatabaseConfigBean">
<property name="skipOracleRecyclebinTables" value="true"/>
</bean>
<bean id="dbUnitDatabaseConnection" class="com.github.springtestdbunit.bean.DatabaseDataSourceConnectionFactoryBean">
<property name="databaseConfig" ref="dbUnitDatabaseConfig"/>
</bean>
NOTE: In most circumstances, the username and password properties should not be set on the DatabaseDataSourceConnectionFactoryBean. These properties will cause DBUnit to start a new transaction and may cause unexpected behaviour.
Several dataset loaders are already available to read from:
- flat XML files, (default)
- CSV files,
- XLS files.
If you need to load data from another source you will need to write your own DataSet loader and configure your tests to use it. Custom loaders must implement the DataSetLoader interface and provide an implementation of the loadDataSet method. The AbstractDataSetLoader is also available and provides a convenient base class for most loaders.
Unlike XML and XLS files, CSV files have a very basic structure which doesn't allow multiple tables to be defined in a single file.
The CsvUrlDataSetLoader therefore relies on a folder containing:
- a
table-ordering.txtfile listing the tables in the order they should be imported, - a dedicated
<table name>.csvfile for each table which has to be imported.
The dataset parameter should therefore point to that folder, not to individual files.
Another point to keep in mind is that the current implementation of DBUnit uses commas as a value separator, not semicolons.
Similarly to CSV files, SQL datasets can only contain one table per file.
The SqlLoaderControlDataSetLoader therefore relies on a folder containing:
- a
tables.lstfile listing the tables in the order they should be imported, - a dedicated
<table name>.ctlfile for each table which has to be imported.
The dataset parameter should therefore point to that folder, not to individual files.
Although the dataset loader is usually defined on the DbUnitConfiguration annotation, it is possible to override it on the following annotations:
DatabaseSetupDatabaseTearDownExpectedDatabase
In some situations you may need to use custom DBUnit DatabaseOperation classes. For example, DBUnit includes org.dbunit.ext.mssql.InsertIdentityOperation for use with Microsoft SQL Server. The DatabaseOperationLookup interface can be used to create your own lookup strategy if you need support custom operations. A MicrosoftSqlDatabaseOperationLookup class is provided to support the aforementioned MSSQL operations.
See above for details of how to configure a test class to use the custom lookup.
A NullPointerException can sometimes occur when loading data from XLS files. This issue is related to how the XLS file was edited. Once content has been submitted into cells and even if these cells are later made blank, the XLS file might still store those empty cells. This will cause the line count for the file to be incorrect, hence the exception.
The solution is quite easy, simply delete the last rows of your file, even if they already seem blank.
You can check the resolution by debugging org.dbunit.dataset.excel.XlsTable.getRowCount() and making sure it returns the correct number of rows.
Have you found an issue? Do you have an idea for an improvement? Feel free to contribute by submitting it on the GitHub project.