Why Unit testing services in NestJS? - Purpose & Use Cases

The Big Idea

What if you could catch bugs instantly without clicking through your app every time?

The Scenario

Imagine you build a NestJS service that handles user data. Every time you change the code, you manually test each function by running the app and clicking through the UI to see if it still works.

The Problem

This manual testing is slow, tiring, and easy to miss bugs. You might forget some cases or introduce errors without noticing. It's like checking every light bulb in a big building by walking room to room instead of having a quick switch test.

The Solution

Unit testing services lets you write small, automatic tests for each function. NestJS testing tools run these tests quickly and show exactly what works or breaks, so you fix problems early without guessing.

Before vs After

✗ Before

Run app -> Click UI -> Check if data saved correctly

✓ After

await expect(service.saveUser(data)).resolves.toEqual(expectedResult)

What It Enables

It enables fast, reliable checks of your service logic anytime you change code, catching bugs before they reach users.

Real Life Example

When adding a new feature to send emails, unit tests verify the email service sends correct messages without needing to send real emails every time.

Key Takeaways

Manual testing is slow and error-prone for service logic.

Unit tests automate checks and catch bugs early.

NestJS provides tools to easily write and run these tests.

Practice

1. What is the main purpose of unit testing a service in NestJS?

easy

A. To verify the service works correctly on its own without dependencies

B. To test the entire application including all modules

C. To check the UI components connected to the service

D. To deploy the service to production automatically

Solution

1. Step 1: Understand unit testing scope

   Unit testing focuses on testing a single piece of code in isolation, here the service alone.

2. Step 2: Identify what unit testing excludes

   It excludes integration with other modules or UI, focusing only on the service logic.

3. Final Answer:

   To verify the service works correctly on its own without dependencies -> Option A

4. Quick Check:

   Unit testing = isolated service check [OK]

Hint: Unit test means testing one piece alone, no dependencies [OK]

Common Mistakes:

• Confusing unit test with integration test
• Thinking unit test covers UI or deployment
• Assuming unit test runs the whole app

2. Which of the following is the correct way to get a service instance in a NestJS unit test?

easy

A. const service = require('MyService');

B. const service = new MyService();

C. const service = module.get<MyService>(MyService);

D. const service = app.getService(MyService);

Solution

1. Step 1: Use NestJS TestingModule

   In unit tests, you create a TestingModule and get the service instance via module.get.

2. Step 2: Identify correct syntax

   The syntax module.get<MyService>(MyService) is the official way to retrieve the service from the test module.

3. Final Answer:

   const service = module.get<MyService>(MyService); -> Option C

4. Quick Check:

   Use module.get to get service instance [OK]

Hint: Always get service from TestingModule with module.get<Type>(Type) [OK]

Common Mistakes:

• Instantiating service directly without module
• Using app.getService which is not a NestJS method
• Using require which is not for NestJS DI

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

const mockData = ['a', 'b'];
const mockRepo = { findAll: () => mockData };

const module = await Test.createTestingModule({
  providers: [
    MyService,
    { provide: MyRepository, useValue: mockRepo },
  ],
}).compile();

const service = module.get<MyService>(MyService);

console.log(service.getData());

medium

A. An error because MyRepository is not mocked

B. undefined

C. [] (empty array)

D. ['a', 'b']

Solution

1. Step 1: Understand mocking of MyRepository

   The mockRepo object replaces MyRepository with a findAll method returning ['a', 'b'].

2. Step 2: Analyze service.getData()

   Assuming getData calls repository.findAll(), it will return ['a', 'b'] from the mock.

3. Final Answer:

   ['a', 'b'] -> Option D

4. Quick Check:

   Mocked repo returns ['a', 'b'] so service returns same [OK]

Hint: Mocked methods return mock values used by service [OK]

Common Mistakes:

• Forgetting to mock repository causing undefined
• Expecting real database call instead of mock
• Assuming empty array without mock data

4. What is the main issue in this test setup code?

const module = await Test.createTestingModule({
  providers: [MyService],
}).compile();

const service = module.get<MyService>(MyService);

jest.spyOn(service, 'findAll').mockReturnValue(['x', 'y']);

medium

A. You cannot spy on service methods before mocking dependencies

B. The service method findAll should be mocked on the dependency, not service

C. jest.spyOn syntax is incorrect here

D. Missing async keyword on test function

Solution

1. Step 1: Identify what to mock in unit tests

   In unit tests, you mock dependencies of the service, not the service's own methods.

2. Step 2: Understand why mocking service method is wrong

   Mocking service.findAll replaces the method under test, defeating the purpose of testing it.

3. Final Answer:

   The service method findAll should be mocked on the dependency, not service -> Option B

4. Quick Check:

   Mock dependencies, not the service itself [OK]

Hint: Mock dependencies, never the service method you test [OK]

Common Mistakes:

• Mocking the service method instead of its dependencies
• Confusing spyOn usage with mocking dependencies
• Ignoring async test function requirements

5. You want to test a service method that calls two different repository methods internally. How should you set up your unit test to isolate the service logic correctly?

hard

A. Mock both repository methods with expected return values and inject them into the test module

B. Call the real repository methods to test the full flow

C. Mock only one repository method and leave the other real

D. Do not mock anything and test the service directly

Solution

1. Step 1: Identify dependencies to mock

   Since the service calls two repository methods, both should be mocked to isolate service logic.

2. Step 2: Setup test module with mocks

   Provide mock implementations for both methods in the test module providers to control their outputs.

3. Step 3: Avoid real calls in unit tests

   Calling real repository methods would make it an integration test, not a unit test.

4. Final Answer:

   Mock both repository methods with expected return values and inject them into the test module -> Option A

5. Quick Check:

   Mock all dependencies to isolate service logic [OK]

Hint: Mock all called dependencies to isolate service logic [OK]

Common Mistakes:

• Mocking only some dependencies causing partial isolation
• Calling real methods making it integration test
• Not injecting mocks properly into test module