Programmers Stack Exchange is a question and answer site for professional programmers interested in conceptual questions about software development. It's 100% free.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This question already has an answer here:

Suppose I have component A,B,C that all put some data into a database. Each component depends on data in the database inserted by the previous component.

How do I test these modules in isolation? I already have tests at the unit level, but I now want to test with the database. I have seen suggestions to make a dump of the database for each step, but this is a messy thing to check in to version control and requires a lot of effort whenever the architecture changes (which where I work translates to "abandon tests!").

I really think that the best approach is to clear my test database before running tests and use some tool to specify test dependencies to ensure correct running order, but this seems to be seriously frowned upon by the opinions I've read. To me it seems the same as using a fixture for each step, except that the fixture is just a cheap way of hiding the dependency.

share|improve this question

marked as duplicate by gnat, Martijn Pieters, 9000, Dynamic, Kilian Foth May 16 '13 at 6:32

This question was marked as an exact duplicate of an existing question.

(1) isn't this called 'integration tests'? (2) how often does your architecture change? how could one trust such changes without tests more that with tests? – 9000 May 15 '13 at 16:30
(1) AFAIK yes, I'm not very familiar with testing terminology. (2) It's a research-oriented project with a small team, so more often than most. It can't be trusted without tests, if a change is needed but it breaks tests, then the tests should be changed. Having to dump a database in a bunch of different states to update the tests is unnecessary. – bcoughlan May 15 '13 at 16:34
look up mocking and dependency injection – ratchet freak May 15 '13 at 21:09

So on a high level you create data sets in code that are applied for each test case.

After each test the database is rolled back either by dropping and recreating or truncating.

You don't mention the framework or language you are using, but many have methods for using fixtures to create data for your tests that are reset during each test cycle.

What you need to do is pick one of these frameworks and setup the core set of data elements you need for your application. Then for each test, the test framework should have a mechanism to populate specific objects (rows) with the data you need for that test scenario. The framework should then reload the database and and apply the default data for the next test. is and example of a fixture framework for Ruby on Rails.

share|improve this answer

Not the answer you're looking for? Browse other questions tagged or ask your own question.