Test one function, with multiple test cases but only use one assert?

Question

I have a class method that takes a list and determines if the list is valid or not according to a function.

I want to test it on three lists that are stored as static variables because they are used in other unit tests later on in the code.

def test__validate(self):
    decoder = Validator()
    slow_valid = Validator.validate(TestValidator.list_slow)
    med_valid = Validator.validate(TestValidator.list_med)
    fast_valid = Validator.validate(TestValidator.list_fast)


    assert slow_valid == True
    assert med_valid == False
    assert fast_valid == False

What is the correct way to remove the multiple assert statements?

Do I defined multiple versions of test__validate or are the multiple assert statements okay from a best practices position?

Answer 1

As proposed by @IanShelvington, the best practise for repeated tests with changed input (and result, in your case) is test parametrization. With pytest, you could do something like this:

import pytest

@pytest.mark.parametrize("val_list, result",
                         [(TestValidator.list_slow, True),
                          (TestValidator.list_med, False),
                          (TestValidator.list_fast, False)])

def test_validate(val_list, result):
    assert Validator().validate(val_list) == result

This gives you the output:

============================= test session starts =============================
...
collecting ... collected 3 items

param_result.py::test_validate[val_list0-True] PASSED                    [ 33%]
param_result.py::test_validate[val_list1-False] PASSED                   [ 66%]
param_result.py::test_validate[val_list2-False] PASSED                   [100%]

============================== 3 passed in 0.04s ==============================

As you can see, this creates 3 separate tests, with the parameters in the name, so a failing test can easily be identified.

If you want customized names of the shown tests, you can provide them using ids:

@pytest.mark.parametrize("val_list, result",
                         [(TestValidator.list_slow, True),
                          (TestValidator.list_med, False),
                          (TestValidator.list_fast, False)],
                         ids=('slow', 'med', 'fast'))
...

This would output:

============================= test session starts =============================
...
param_result.py::test_validate[slow] PASSED                              [ 33%]
param_result.py::test_validate[med] PASSED                               [ 66%]
param_result.py::test_validate[fast] PASSED                              [100%]

============================== 3 passed in 0.06s ==============================

Answer 2

Since Python 3.4, You can do this with subTest in unittest.

import unittest

class TestValidation(unittest.TestCase):
    def test_validate(self):
        decoder = Validator()
        validation_list = [
            ("slow", Validator.validate(TestValidator.list_slow), True),
            ("med", Validator.validate(TestValidator.list_med), False),
            ("fast", Validator.validate(TestValidator.list_fast), False),
        ]

        for label, actual_output, expected_output in validation_list:
            with self.subTest(label=label):
                self.assertEqual(expected_output, actual_output)

The extra label field in the three tuples is unnecessary but will make the error much easier to debug in case any of the tests fail. The label text will appear in the SubTest error message when any one of the three tests fails, allowing you to diagnose which of the three failed.

Note: I noticed after answering the question was tagged pytest. Anyways, unittest users may come across and hopefully find this answer helpful.