Testing with database in FastAPI - Performance & Optimization

Performance: Testing with database

MEDIUM IMPACT

This concept affects the speed of test execution and the responsiveness of the development feedback loop.

Running automated tests that require database access

FastAPI

import pytest
from fastapi.testclient import TestClient
from sqlalchemy import create_engine
from sqlalchemy.orm import sessionmaker
from myapp.main import app
from myapp.database import Base, get_db

# Use in-memory SQLite for tests
SQLALCHEMY_DATABASE_URL = 'sqlite:///:memory:'
engine = create_engine(SQLALCHEMY_DATABASE_URL, connect_args={'check_same_thread': False})
TestingSessionLocal = sessionmaker(autocommit=False, autoflush=False, bind=engine)

Base.metadata.create_all(bind=engine)

def override_get_db():
    db = TestingSessionLocal()
    try:
        yield db
    finally:
        db.close()

app.dependency_overrides[get_db] = override_get_db
client = TestClient(app)

def test_create_item():
    response = client.post('/items/', json={'name': 'test'})
    assert response.status_code == 200
    # This test uses a fast in-memory database

Tests run against a lightweight in-memory database, avoiding network and disk delays.

📈 Performance GainReduces test runtime by 70-90%; enables faster feedback and parallel test execution

Running automated tests that require database access

FastAPI

import pytest
from fastapi.testclient import TestClient
from myapp.main import app

client = TestClient(app)

def test_create_item():
    response = client.post('/items/', json={'name': 'test'})
    assert response.status_code == 200
    # This test uses the real database connection

Tests connect to the real database causing slow setup, teardown, and possible data pollution.

📉 Performance CostBlocks test execution for 100-500ms per test; slows CI pipelines significantly

Performance Comparison

Pattern	Test Runtime	I/O Operations	Resource Usage	Verdict
Real database connection	Slow (100-500ms per test)	High (network + disk)	High (DB server resources)	[X] Bad
In-memory SQLite database	Fast (10-50ms per test)	Low (memory only)	Low (no external resources)	[OK] Good

Rendering Pipeline

Testing with a real database triggers network and disk I/O, increasing test runtime and blocking the event loop. Using an in-memory database reduces these costs by keeping data in RAM and avoiding external calls.

→Test Execution

→I/O Operations

→Event Loop Blocking

⚠️ BottleneckI/O Operations with real database connections

Optimization Tips

1Avoid using real databases in automated tests to speed up test execution.

2Use in-memory or mocked databases to reduce I/O and resource usage during tests.

3Measure test durations to identify slow database-dependent tests and optimize them.

Performance Quiz - 3 Questions

What is the main performance drawback of using a real database in FastAPI tests?

A Tests cause browser rendering delays

B Tests use too much CPU for calculations

C Tests run slower due to network and disk I/O

D Tests increase CSS paint cost

DevTools: Terminal / CI logs

How to check: Measure test suite runtime before and after switching to in-memory DB; use pytest --durations to find slow tests

What to look for: Significant reduction in test execution time and less waiting on database setup/teardown

Practice

1. Why should you use a separate test database when testing FastAPI applications that interact with a database?

easy

A. To skip database setup during tests

B. To make tests run slower and more realistic

C. To avoid changing real data and keep tests isolated

D. To share data between tests and production

Solution

1. Step 1: Understand test isolation

   Tests should not affect real data to avoid accidental data loss or corruption.

2. Step 2: Use separate test database

   Using a separate database keeps test data separate and allows safe repeated testing.

3. Final Answer:

   To avoid changing real data and keep tests isolated -> Option C

4. Quick Check:

   Separate test database = safe isolated tests [OK]

Hint: Always isolate test data from real data [OK]

Common Mistakes:

• Using production database for tests
• Not cleaning test data after tests
• Assuming tests don't affect real data

2. Which of the following is the correct way to override the database dependency in FastAPI tests?

easy

A. app.replace_dependency(get_db, override_get_db)

B. app.override_dependency(get_db, override_get_db)

C. app.set_dependency(get_db, override_get_db)

D. app.dependency_overrides[get_db] = override_get_db

Solution

1. Step 1: Recall FastAPI dependency override syntax

   FastAPI uses the dictionary app.dependency_overrides to replace dependencies during tests.

2. Step 2: Match correct syntax

   Assigning override_get_db function to app.dependency_overrides[get_db] is the correct method.

3. Final Answer:

   app.dependency_overrides[get_db] = override_get_db -> Option D

4. Quick Check:

   Override dependencies via app.dependency_overrides dict [OK]

Hint: Use app.dependency_overrides dictionary to override [OK]

Common Mistakes:

• Using non-existent methods like set_dependency
• Trying to override dependencies without assignment
• Confusing syntax with other frameworks

3. Given this test code snippet, what will be the output of response.json()?

def test_create_item(client):
    response = client.post('/items/', json={'name': 'Book', 'price': 10})
    return response.json()

medium

A. {'id': 1, 'name': 'Book', 'price': 10}

B. {'name': 'Book', 'price': 10}

C. An error because POST is not allowed

D. Empty response {}

Solution

1. Step 1: Understand POST /items/ behavior

   Assuming the endpoint creates an item and returns it with an assigned id.

2. Step 2: Analyze response.json()

   The response should include the new item's id plus the sent data.

3. Final Answer:

   {'id': 1, 'name': 'Book', 'price': 10} -> Option A

4. Quick Check:

   POST creates item and returns it with id [OK]

Hint: POST usually returns created object with id [OK]

Common Mistakes:

• Expecting only sent data without id
• Assuming POST is disallowed
• Ignoring response content

4. What is the main issue with this test code snippet?

def override_get_db():
    db = SessionLocal()
    try:
        yield db
    finally:
        db.close()

app.dependency_overrides[get_db] = override_get_db

def test_read_items(client):
    response = client.get('/items/')
    assert response.status_code == 200
    assert response.json() == []

medium

A. The dependency override is missing

B. The test database session is not properly created or isolated

C. The test client is not defined

D. The test expects wrong status code

Solution

1. Step 1: Check override_get_db function

   It uses SessionLocal but does not ensure a separate test database or transaction rollback.

2. Step 2: Identify test isolation problem

   Without a test database or rollback, tests may affect real data or each other.

3. Final Answer:

   The test database session is not properly created or isolated -> Option B

4. Quick Check:

   Test DB session must be isolated and separate [OK]

Hint: Ensure test DB session is isolated and rolled back [OK]

Common Mistakes:

• Not using a test database
• Not overriding dependencies correctly
• Assuming client is auto-defined

5. You want to write a test that creates an item, then reads it back from the test database in FastAPI. Which approach correctly ensures the test uses a separate database and cleans up after itself?

hard

A. Use a test database URL, override get_db to yield a session from it, and rollback after each test

B. Use the production database URL and delete created items manually after test

C. Use an in-memory list to mock database and skip actual DB calls

D. Run tests without overriding get_db and rely on default DB

Solution

1. Step 1: Use separate test database URL

   This prevents tests from affecting production data.

2. Step 2: Override get_db to yield test session and rollback

   This ensures each test runs isolated and cleans up changes automatically.

3. Final Answer:

   Use a test database URL, override get_db to yield a session from it, and rollback after each test -> Option A

4. Quick Check:

   Separate DB + override + rollback = clean isolated tests [OK]

Hint: Always isolate tests with separate DB and rollback [OK]

Common Mistakes:

• Using production DB for tests
• Not cleaning test data after tests
• Skipping dependency overrides