The Best New Feature in unittest You Didn’t Know You Need

For a better reading experience, check out this article on my website.

From time to time I like to read documentation of modules I think I know well. The python documentation is not a pleasant read but sometimes you strike a gem.

Same thing but slightly different

Distinguishing Test Iterations

Let’s start with a simple function to check if a number is even

def is_even(n):return n % 2 == 0

And a simple test

class TestIsEven(TestCase):

def test\_should\_be\_even(self):  
    self.assertTrue(is\_even(2))

Nice, let’s add some more cases:

class TestIsEven(TestCase):...

def test\_zero\_should\_be\_even(self):  
    self.assertTrue(is\_even(0))

def test\_negative\_should\_be\_even(self):  
    self.assertTrue(is\_even(-2))

This is a simple example and we copied code three times. Let’s try to do better by writing a loop to iterate values we expect to be even:

class TestIsEven(TestCase):

def test\_should\_all\_be\_even(self):  
    for n in (2, 0, -2, 11):  
        self.assertTrue(is\_even(n))

This is starting to look more elegant, so what is the problem? I added an odd value, 11, to fail the test. Let’s run the test and see what it looks like:

F===================================================FAIL: test_should_all_be_even (__main__.TestIsEven)— — —— — — — — — — — — — — — — — — — — — — — — — —

Traceback (most recent call last):

File “subtest.py”, line 18, in test_should_all_be_evenself.assertTrue(is_even(n))AssertionError: False is not true

It failed as expected, but which value failed?

Enter subTest

In python 3.4 there is a new feature called subTest. Lets see it in action:

class TestIsEven(TestCase):

def test\_should\_all\_be\_even(self):  
    for n in (0, 4, -2, 11):  
        **with self.subTest(n=n):**                self.assertTrue(is\_even(n))

Running this test produces the following output:

F==========================================================FAIL: test_should_all_be_even (__main__.TestIsEven) (n=11)— — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Traceback (most recent call last):

File “subtest.py”, line 23, in test_should_all_be_evenself.assertTrue(is_even(n))AssertionError: False is not true

So which value failed? 11! It’s in the title.

How multiple failures look like?

F===========================================================FAIL: test_should_all_be_even (__main__.TestIsEven) (n=3)— — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Traceback (most recent call last):

File “subtest.py”, line 23, in test_should_all_be_evenself.assertTrue(is_even(n))AssertionError: False is not true

==========================================================FAIL: test_should_all_be_even (__main__.TestIsEven) (n=5)— — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Traceback (most recent call last):

File “subtest.py”, line 23, in test_should_all_be_evenself.assertTrue(is_even(n))AssertionError: False is not true

==========================================================FAIL: test_should_all_be_even (__main__.TestIsEven) (n=11)— — — — — — — — — — — — — — — — — — — — — — — — — — — — —

Traceback (most recent call last):

File “subtest.py”, line 23, in test_should_all_be_evenself.assertTrue(is_even(n))AssertionError: False is not true

Exactly as if we wrote three separate test cases.

Profit!