Test Parametrization from Files¶

The parametrize_from_fixture decorator allows you to parametrize tests using data from external files. This is particularly useful for data-driven testing where you want to separate test logic from test data.

Warning

This decorator expects the data files to be in the tests/fixtures/ directory by default – this can not be changed for now.

Quick Start¶

Given a CSV file with test data:

id,a,b,c
add_positive,1,2,3
add_negative,1,-1,0
add_zero,5,0,5

You can parametrize a test function like this:

from pytest_fixtures_fixtures import parametrize_from_fixture

@parametrize_from_fixture("data.csv")
def test_addition(a, b, c):
    assert int(a) + int(b) == int(c)

This will expand into three separate tests: - test_addition[add_positive] - test_addition[add_negative]
- test_addition[add_zero]

Supported File Formats¶

CSV Files¶

CSV files use the first row as headers to define parameter names.

id,username,password,expected_status
valid_login,alice,secret123,200
invalid_password,alice,wrong,401
missing_user,bob,anything,404

JSON Files¶

JSON files should contain an array of objects:

[
    {"id": "valid_request", "method": "GET", "endpoint": "/users", "status": 200},
    {"id": "invalid_endpoint", "method": "GET", "endpoint": "/invalid", "status": 404}
]

YAML Files¶

YAML files should contain a list of dictionaries:

- id: config_dev
  environment: development
  debug: true
  database: sqlite
- id: config_prod  
  environment: production
  debug: false
  database: postgresql

JSONL Files¶

JSONL files contain one JSON object per line:

{"id": "user_admin", "role": "admin", "permissions": ["read", "write", "delete"]}
{"id": "user_guest", "role": "guest", "permissions": ["read"]}

Custom Test IDs¶

The decorator automatically extracts test IDs from an id field in your data files. When present, these IDs are used to make test identification more meaningful.

Without IDs (default pytest behavior)¶

a,b,c
1,2,3
3,4,7

Results in: test_addition[1-2-3], test_addition[3-4-7]

With IDs (custom meaningful names)¶

id,a,b,c
simple_addition,1,2,3
with_larger_numbers,3,4,7

Results in: test_addition[simple_addition], test_addition[with_larger_numbers]

Advanced Options¶

Custom ID Field¶

You can specify a different field name for test IDs:

@parametrize_from_fixture("test_data.csv", id_field="test_name")
def test_with_custom_id_field(param1, param2):
    # Uses 'test_name' column for test IDs
    pass

Disable ID Extraction¶

Set id_field=None to disable automatic ID extraction:

@parametrize_from_fixture("test_data.csv", id_field=None)
def test_with_default_ids(param1, param2):
    # Uses pytest's default ID generation
    pass

User-Provided IDs Override¶

User-provided IDs always take precedence over file-based IDs:

@parametrize_from_fixture("test_data.csv", ids=["custom_1", "custom_2"])
def test_with_override_ids(param1, param2):
    # Uses custom_1, custom_2 regardless of file content
    pass

Explicit File Format¶

Override automatic format detection:

@parametrize_from_fixture("data.txt", file_format="csv")
def test_explicit_format(param1, param2):
    # Treats data.txt as CSV regardless of extension
    pass

Custom Fixtures Directory¶

Override the default fixtures directory:

@parametrize_from_fixture("test_data.csv", fixtures_dir="custom/fixtures")
def test_custom_dir(param1, param2):
    # Looks in custom/fixtures/ instead of tests/fixtures/
    pass

Integration with pytest.mark.parametrize¶

The decorator supports all pytest.mark.parametrize options:

@parametrize_from_fixture(
    "test_data.csv",
    indirect=True,
    scope="class"
)
def test_with_pytest_options(param1, param2):
    # Uses indirect=True and scope="class"
    pass

Error Handling¶

The decorator provides clear error messages for common issues:

File not found: FileNotFoundError: Fixture test_data.csv does not exist
Invalid format: ValueError: Cannot auto-detect format for file: data.unknown
Inconsistent data: ValueError: All items in JSON file must have the same keys
Empty data: ValueError: CSV file has no data rows

Best Practices¶

Use descriptive test IDs that make it easy to identify failing tests
Keep data files focused - one file per logical test group
Use consistent field names across related test files
Include edge cases in your test data
Validate your test data - ensure it's well-formed and complete

Migration from Manual Parametrization¶

Before:¶

@pytest.mark.parametrize("a,b,expected", [
    (1, 2, 3),
    (3, 4, 7), 
    (5, 6, 11),
])
def test_addition(a, b, expected):
    assert a + b == expected

After:¶

Create addition_tests.csv:

id,a,b,expected
case_1,1,2,3
case_2,3,4,7
case_3,5,6,11

Update test:

@parametrize_from_fixture("addition_tests.csv")
def test_addition(a, b, expected):
    assert int(a) + int(b) == int(expected)

Warning

Note that CSV files will be read as strings by default, so you will need to convert them to integers if you want to use them in your test.

Benefits:¶

External data is easier to maintain and review
Non-developers can contribute test cases
Test data can be generated programmatically
Better separation of test logic and test data
More descriptive test IDs for better debugging