DbSetup allows populating a database before executing automated integration tests
(typically, DAO/Repository automated tests). Although DBUnit, which is a great project, allows doing the same thing
and much more, it's also harder to use and setup. And in our experience, in 98% of the cases, DBUnit is only used
to pre-populate a database before executing every test method. This is the task on which DbSetup concentrates.
The philosophy of DbSetup is that DAO tests should not have to setup the database, execute tests, and then
remove everything from the database. Instead, a single setup method should be used to delete everything from the
database (whatever the previous test put in it, or the initial state of the database tables), and then populate it
with the data necessary to execute the test.
Another design choice of DbSetup is to provide an easy to use and simple Java API to populate the database, rather than
loading data from external XML files. Using a Java API has several advantages:
- It allows using real Java types as data (longs, enums, etc.)
- It allows defining default values, looping to generate several similar rows, storing data sets in variables
or factorizing their creation using reusable methods
- It allows viewing the data sets easily, without having to open external files, by storing the data set
directly into the test class, or by navigating through classes and methods using the IDE shortcuts.
- For more complex situations, like cyclic referential integrity constraints between rows, the Java API
allows easily integrating SQL statements into the sequence of operations to execute to pre-populate the
database. These SQL statements can, for example, disable constraints and re-enable them, or update rows
after their insertion.
The
DbSetup class and the
DbSetupTracker classes
are the main entry points to the API that an automated test will use. Look at their javadoc below for example usage.