Working with Data
Tests should not affect each other. That’s a rule of thumb. When tests interact with a database,
they may change the data inside it, which would eventually lead to data inconsistency.
A test may try to insert a record that has already been inserted, or retrieve a deleted record.
To avoid test failures, the database should be brought back to its initial state before each test.
Codeception has different methods and approaches to get your data cleaned.
This chapter summarizes all of the notices on clean-ups from the previous chapters
and suggests the best strategies of how to choose data storage backends.
When we decide to clean up a database, we should make this cleaning as fast as possible. Tests should always run fast.
Rebuilding the database from scratch is not the best way, but might be the only one. In any case,
you should use a special test database for testing. Do not ever run tests on development or production databases!
Codeception has a
Db module, which takes on most of the tasks of database interaction.
By default it will try to repopulate the database from a dump and clean it up after each test.
This module expects a database dump in SQL format. It’s already prepared for configuration in
Use module parameters
to set the database credentials from environment variables or from application configuration files.
After you enable this module in your test suite, it will automatically populate the database from a dump
and repopulate it on each test run. These settings can be changed through the
which may be set to
A full database clean-up can be painfully slow if you use large database dumps. It is recommended to do more data testing
on the functional and integration levels, this way you can get performance bonuses from using ORM.
In acceptance tests, your tests are interacting with the application through a web server. This means that the test
and the application work with the same database. You should provide the same credentials in the Db module
that your application uses, then you can access the database for assertions (
and to perform automatic clean-ups.
The Db module provides actions to create and verify data inside a database.
If you want to create a special database record for one test,
you can use
haveInDatabase method of
haveInDatabase inserts a row with the provided values into the database.
All added records will be deleted at the end of the test.
If you want to check that a table record was created
See the module reference for other methods you can use for database testing.
There are also modules for MongoDb,
and Memcache which behave in a similar manner.
If the database clean-up takes too long, you can follow a different strategy: create new data for each test.
This way, the only problem you might face is duplication of data records.
Sequence was created to solve this.
It provides the
sq() function which generates unique suffixes for creating data in tests.
Your application is most likely using object-relational mapping (ORM) to work with the database. In this case,
Codeception allows you to use the ORM methods to work with the database, instead of accessing the database directly.
This way you can work with models and entities of a domain, and not on tables and rows.
By using ORM in functional and integration tests, you can also improve performance of your tests.
Instead of cleaning up the database after each test, the ORM module will wrap all the database actions into transactions
and roll it back at the end. This way, no actual data will be written to the database.
This clean-up strategy is enabled by default,
you can disable it by setting
cleanup: false in the configuration of any ORM module.
Popular frameworks like Laravel, Yii, and Phalcon include an ActiveRecord data layer by default.
Because of this tight integration, you just need to enable the framework module, and use its configuration for database access.
Corresponding framework modules provide similar methods for ORM access:
They allow you to create and check data by model name and field names in the model. Here is the example in Laravel:
Laravel5 module also provides `haveModel`, `makeModel` methods which use factories to generate models with fake data.
If you want to use ORM for integration testing only, you should enable the framework module with only the
ORM part enabled:
This way no web actions will be added to
If you want to use ORM to work with data inside acceptance tests, you should also include only the ORM part of a module.
Please note that inside acceptance tests, web applications work inside a webserver, so any test data can’t be cleaned up
by rolling back transactions. You will need to disable cleaning up,
and use the
Db module to clean the database up between tests. Here is a sample config:
Doctrine is also a popular ORM, unlike some others it implements the DataMapper pattern and is not bound to any framework.
The Doctrine2 module requires an
EntityManager instance to work with.
It can be obtained from a Symfony framework or Zend Framework (configured with Doctrine):
If no framework is used with Doctrine you should provide the
with a valid callback to a function which returns an
Doctrine2 also provides methods to create and check data:
Preparing data for testing is a very creative, although boring, task. If you create a record,
you need to fill in all the fields of the model. It is much easier to use Faker
for this task, which is more effective to set up data generation rules for models.
Such a set of rules is called factories
and are provided by the DataFactory module.
Once configured, it can create records with ease:
Created records will be deleted at the end of a test.
The DataFactory module only works with ORM, so it requires one of the ORM modules to be enabled:
DataFactory provides a powerful solution for managing data in integration/functional/acceptance tests.
Read the full reference to learn how to set this module up.
Codeception also assists the developer when dealing with data. Tools for database population
and cleaning up are bundled within the
Db module. If you use ORM, you can use one of the provided framework modules
to operate with database through a data abstraction layer, and use the DataFactory module to generate new records with ease.