Ordering tests for parallel safety in PyTest

Introduction

Sometimes tests run at the same time can interfere with each other. Ordering tests helps keep them safe and reliable.

When tests share resources like files or databases and might change data at the same time.

When one test depends on the result or setup of another test.

When running tests in parallel causes failures that don't happen when running them one by one.

When you want to speed up testing but keep tests from breaking each other.

Syntax

PyTest

import pytest

@pytest.mark.order(1)
def test_first():
    pass

@pytest.mark.order(2)
def test_second():
    pass

Use @pytest.mark.order(n) to set the order where n is a number.

Lower numbers run first. Tests without order run after ordered tests.

Examples

This example runs test_setup first, then test_main, and finally test_cleanup.

PyTest

import pytest

@pytest.mark.order(1)
def test_setup():
    assert True

@pytest.mark.order(3)
def test_cleanup():
    assert True

@pytest.mark.order(2)
def test_main():
    assert True

Tests with no order run after tests with order numbers.

PyTest

import pytest

def test_no_order():
    assert True

@pytest.mark.order(1)
def test_ordered():
    assert True

Sample Program

This test suite uses ordering to safely add, check, and clear a shared list. Running in order prevents errors from parallel access.

PyTest

import pytest

shared_resource = []

@pytest.mark.order(1)
def test_add_item():
    shared_resource.append('item')
    assert 'item' in shared_resource

@pytest.mark.order(2)
def test_check_item():
    assert 'item' in shared_resource

@pytest.mark.order(3)
def test_clear_resource():
    shared_resource.clear()
    assert shared_resource == []

Important Notes

Ordering tests can help but avoid making tests depend too much on each other.

Use fixtures to share setup and teardown instead of relying only on order.

Parallel test runners like pytest-xdist may still run tests in parallel; ordering helps but does not guarantee isolation.

Summary

Ordering tests helps avoid conflicts when tests share resources.

Use @pytest.mark.order(n) to control test run order.

Keep tests independent when possible for easier maintenance.

Practice

1. Why is it important to order tests when running pytest in parallel?

easy

A. To make tests run slower

B. To avoid conflicts when tests share resources

C. To increase the number of tests

D. To skip tests automatically

Solution

1. Step 1: Understand test resource sharing

   Tests that share resources like files or databases can interfere with each other if run at the same time.

2. Step 2: Recognize the role of ordering

   Ordering tests ensures they run in a sequence that prevents conflicts and errors.

3. Final Answer:

   To avoid conflicts when tests share resources -> Option B

4. Quick Check:

   Ordering prevents resource conflicts [OK]

Hint: Order tests to prevent shared resource conflicts [OK]

Common Mistakes:

• Thinking ordering slows tests down
• Believing ordering increases test count
• Assuming ordering skips tests

2. Which of the following is the correct way to order a test to run third using pytest?

easy

A. @pytest.mark.order(3)

B. @pytest.order(3)

C. @pytest.mark.run(3)

D. @pytest.order_mark(3)

Solution

1. Step 1: Recall pytest ordering syntax

   The correct decorator to order tests is @pytest.mark.order(n) where n is the order number.

2. Step 2: Match the syntax to options

   Only @pytest.mark.order(3) uses the correct decorator @pytest.mark.order(3).

3. Final Answer:

   @pytest.mark.order(3) -> Option A

4. Quick Check:

   Use @pytest.mark.order(n) for ordering [OK]

Hint: Use @pytest.mark.order(n) to set test order [OK]

Common Mistakes:

• Using @pytest.order instead of @pytest.mark.order
• Confusing order with run decorators
• Misspelling the decorator name

3. Given these two tests, what will be the order of execution when run with pytest in parallel with ordering?

import pytest

@pytest.mark.order(2)
def test_second():
    assert True

@pytest.mark.order(1)
def test_first():
    assert True

medium

A. Tests run in random order

B. test_second runs before test_first

C. test_first runs before test_second

D. Tests fail due to ordering conflict

Solution

1. Step 1: Identify order markers

   test_first has order 1, test_second has order 2.

2. Step 2: Understand execution order

   Lower order numbers run before higher ones, so test_first runs before test_second.

3. Final Answer:

   test_first runs before test_second -> Option C

4. Quick Check:

   Lower order number runs first [OK]

Hint: Lower order number runs first in pytest [OK]

Common Mistakes:

• Assuming higher order runs first
• Thinking tests run randomly despite order
• Believing ordering causes test failure

4. You have two tests that share a database. You want to run them in parallel safely. Which of these is a problem in the code below?

import pytest

@pytest.mark.order(1)
def test_write_db():
    # writes data
    assert True

@pytest.mark.order(2)
def test_read_db():
    # reads data
    assert True

medium

A. Tests must have the same order number

B. The order decorators are incorrect syntax

C. Tests are missing assert statements

D. Tests are ordered but may still run in parallel causing conflicts

Solution

1. Step 1: Check ordering usage

   Tests use correct order decorators, so syntax is fine.

2. Step 2: Understand parallel execution impact

   Even with order, if tests run truly in parallel (e.g., with pytest-xdist), they may overlap and cause conflicts.

3. Final Answer:

   Tests are ordered but may still run in parallel causing conflicts -> Option D

4. Quick Check:

   Ordering alone doesn't guarantee parallel safety [OK]

Hint: Ordering doesn't prevent parallel overlap without locks [OK]

Common Mistakes:

• Thinking order decorators fix parallel conflicts
• Believing same order number is required
• Ignoring assert statements importance

5. You have three tests that modify a shared file. To run them safely in parallel, you want to order them and ensure no overlap. Which approach below best achieves this?

hard

A. Use @pytest.mark.order to run tests sequentially and add file locks

B. Remove order decorators and run all tests in parallel without locks

C. Use @pytest.mark.order with the same order number for all tests

D. Run tests without ordering but add random sleep delays

Solution

1. Step 1: Understand test resource sharing

   Tests modifying the same file can cause conflicts if run simultaneously.

2. Step 2: Combine ordering with locking

   Ordering ensures sequence, and file locks prevent overlap during execution.

3. Step 3: Evaluate options

   Use @pytest.mark.order to run tests sequentially and add file locks uses both ordering and locks, which is the safest approach.

4. Final Answer:

   Use @pytest.mark.order to run tests sequentially and add file locks -> Option A

5. Quick Check:

   Order plus locks ensure parallel safety [OK]

Hint: Combine order and locks for safe parallel file tests [OK]

Common Mistakes:

• Relying on order alone without locks
• Using same order number causing race conditions
• Adding random delays instead of proper synchronization