dataz

dataZ provides testing support for Datastores.

A typical Datastore is a database, either SQL or NoSql. But actually, any kind of (persistence) storage or with system with side effects like a data base could be a Datastore.

If you have to test a module which is using a Datastore, you have are facing some of these typical issues:

• Tests are slow.
• Tests are brittle.
• Tests are integration tests.
• Tests could not executed concurrent.
• How should I provide test data (test fixture), so that the tests are repeatable and independent.
• Test fixtures are hard to maintain: Just add a column to a table ...
• A complex query to a database is worth it's own unit test (even this kind of test some people won't call this an unit test at all).
• Mocks are ... (your choice :-))

More issues are

• The tests should be fast and so you decide to use a different database then the productive one (for example h2). But ...
• ... the tests should also be executed on the productive database.
• You have more then one datastore. For example: A single master database and for each tenant a different database instance.
• You have different types of datastores. For example: MySql and neo4j.
• You are making a product which is based on JPA, but every test should run also on any supported database.
• Performance tests with the productive database.
• and even more trouble...

Brief: Writing tests are a pain in the ass.

Everything which makes trouble, people try to avoid it. Another point is that is hard to apply TDD if a Datastore is part of your test.

Design Goals

So dataZ makes testing with Datastores a pleasure (again?).

1. The general design goal is to provide a framework which works for any kind of Datastore.
2. It should close the gap between unit and integration tests.
3. Having a open and extendable framework which is adaptable to any specific needs of your project.

Documentation

Under construction

dataZ - doc

Build

TODO: Deploy to ???.

Datastore implementations

Currently are two Datastore implementations available:

• Neo4J
• Relational DB

Example

TODO: Example repository

package com.company.module.my;

import org.failearly.dataz.junit4.DataSetDriver;
import org.failearly.dataz.DataSet;

// JUnit imports omitted for brevity


@DataSet // <<<<<< Expects data (setup) resource /com/company/module/my/MyTest.setup >>>>>>
public class MyTest {
    @Rule public final TestRule _dataZ_Driver = DataSetDriver.createDataSetDriver(this); // <<<< This driver which applies the data resources on a datastores

    @Test
    public void testUsingClassDataSet() {
        // Running your test against data resource /com/company/module/my/MyTest.setup
    }

    @Test @DataSet // <<<<<< Expects data (setup) resource /com/company/module/my/MyTest-testUsingMethodDataSet.setup >>>>>>
    public void testUsingMethodDataSet() {
        // Running your test against data resources 
		//            /com/company/module/my/MyTest.setup 
		//    and(!!) /com/company/module/my/MyTest-testUsingMethodDataSet.setup
    }
}

Another possibility is to use a JUnit base class (AbstractDataSetTest) provided by dataZ.

package com.company.module.my;

import org.failearly.dataz.junit4.AbstractDataSetTest;
import org.failearly.dataz.DataSet;

// JUnit imports omitted for brevity

@DataSet
public class MyTest extends AbstractDataSetTest /* <<< Use this instead of TestRule >>>> */ {
    @Test
    public void testUsingClassDataSet() {
        // ...
    }

    @Test @DataSet
    public void testUsingMethodDataSet() {
        // ...
    }
}

Issues

If you want to report any issue or feature request, please do it here.

Social

You can found me on

Twitter
LinkedIn

License

Eclipse Public License - v2.0