Test Parametrization in Pytest: Running Tests with Multiple Inputs

Parameterized testing is crucial for enhancing test coverage without writing numerous test functions for scenarios that only differ by their inputs and expected outputs. This approach not only saves time but also increases the clarity and maintainability of test suites by:

• Reducing code duplication and the potential for errors in test code;
• Enabling easy addition of new test cases by simply adding parameters;
• Facilitating comprehensive testing across a wide range of inputs.

Let's consider a simple function that calculates the average of two numbers. There can be a logical error if to forget to put braces. We'll use @pytest.mark.parametrize to test this function with several pairs of numbers.

python
Kopier kode
99
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
# Function to test
def calculate_average(num1, num2):
    return (num1 + num2) / 2

# Test function using pytest's parametrize
import pytest

@pytest.mark.parametrize(
    "num1, num2, expected",
    [
        (10, 20, 15),
        (20, 30, 25),
        (5, 5, 5)
    ])
def test_calculate_average(num1, num2, expected):
    assert (calculate_average(num1, num2) == expected), f"Average of {num1} and {num2} should be equal to {(num1 + num2) / 2}"

This example demonstrates how to run test_calculate_average with three different sets of numbers, checking if the function correctly computes their average.

Syntax with id

The id parameter is used to assign an identifier to each set of inputs, which helps to identify tests more easily in the test output:

python
Kopier kode
9
1
2
3
4
5
6
7
 @pytest.mark.parametrize("num1, num2, expected", [
    pytest.param(10, 20, 15, id="integers"),
    pytest.param(20, 30, 25, id="more integers"),
    pytest.param(5, 5, 5, id="equal numbers")
])
def test_calculate_average_with_ids(num1, num2, expected):
    assert calculate_average(num1, num2) == expected

Using pytest.param allows for more granular control over each test case, including setting individual IDs directly.

Syntax with ids

You can also provide a unique ID for each test case using ids, which makes test reports more readable:

python
Kopier kode
9
1
2
3
4
5
6
7
@pytest.mark.parametrize("num1, num2, expected", [
    (10, 20, 15),
    (20, 30, 25),
    (5, 5, 5)
], ids=["integers", "more integers", "equal numbers"])
def test_calculate_average_with_ids(num1, num2, expected):
    assert calculate_average(num1, num2) == expected

In this code, each set of parameters has an associated ID that describes the test case, which will appear in the test output.

Here's what the output might look like:

python
Kopier kode
99
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
===================== test session starts ======================
collected 3 items

test_example.py::test_calculate_average_with_ids[integers] FAILED
test_example.py::test_calculate_average_with_ids[more integers] PASSED
test_example.py::test_calculate_average_with_ids[equal numbers] PASSED

=========================== FAILURES ===========================
________________ test_calculate_average_with_ids[integers] ________________

num1 = 10, num2 = 20, expected = 15

    def test_calculate_average_with_ids(num1, num2, expected):
>       assert calculate_average(num1, num2) == expected
E       assert 20 == 15 
E        +  where 20 = calculate_average(10, 20)

test_example.py:10: AssertionError
================== 1 failed, 2 passed in 0.23s ==================