Why Discriminated union state machines in Typescript? - Purpose & Use Cases

The Big Idea

What if your program could never get confused about what state it's in?

The Scenario

Imagine you are building a program that handles different states, like a traffic light or a login process, and you try to manage each state with separate variables and lots of if-else checks scattered everywhere.

The Problem

This manual way quickly becomes confusing and error-prone. You might forget to check a state, mix up conditions, or write repetitive code that is hard to read and maintain.

The Solution

Discriminated union state machines let you define all possible states clearly in one place with unique tags. This makes your code safer, easier to understand, and TypeScript helps catch mistakes before running your program.

Before vs After

✗ Before

if (state === 'loading') { /* do loading */ } else if (state === 'error') { /* handle error */ } else if (state === 'success') { /* show data */ }

✓ After

type State = { type: 'loading' } | { type: 'error'; message: string } | { type: 'success'; data: string };
switch(state.type) {
  case 'loading': /* do loading */; break;
  case 'error': /* handle error */; break;
  case 'success': /* show data */; break;
}

What It Enables

You can build clear, reliable programs that handle complex states without confusion or bugs.

Real Life Example

Think of a music player app that can be stopped, playing, or paused. Using discriminated unions, you can easily manage what happens in each state and avoid mistakes like trying to pause when already stopped.

Key Takeaways

Manual state checks get messy and error-prone.

Discriminated unions group states with unique tags for clarity.

TypeScript helps catch errors and makes code easier to maintain.

Practice

1.

What is the main purpose of using a discriminated union in TypeScript state machines?

easy

A. To enable dynamic addition of states at runtime without type errors.

B. To clearly define different states with a common type property for safe type checking.

C. To allow any property names in the state objects without restrictions.

D. To automatically generate UI components based on state names.

Solution

1. Step 1: Understand discriminated unions

   Discriminated unions use a common type property to distinguish between different object shapes.

2. Step 2: Identify the purpose in state machines

   This common property allows TypeScript to safely narrow the state type and catch errors.

3. Final Answer:

   To clearly define different states with a common type property for safe type checking. -> Option B

4. Quick Check:

   Discriminated union = safe state typing [OK]

Hint: Look for the common type property in states [OK]

Common Mistakes:

• Thinking discriminated unions allow any property names freely
• Assuming states can be added dynamically without type errors
• Confusing discriminated unions with UI generation

2.

Which of the following is the correct syntax to define a discriminated union type for a state machine with states loading and success?

type State = ?

easy

A. { status: 'loading' } | { status: 'success'; data: string }

B. { type: 'loading'; data: string } | { type: 'success' }

C. { type: 'loading' } | { type: 'success'; data: string }

D. { state: 'loading' } | { state: 'success'; data: string }

Solution

1. Step 1: Identify the discriminant property

   The discriminated union requires a common property named type to distinguish states.

2. Step 2: Match the correct state shapes

   Loading state has only type: 'loading', success state has type: 'success' and extra data.

3. Final Answer:

   { type: 'loading' } | { type: 'success'; data: string } -> Option C

4. Quick Check:

   Common type property with correct fields = B [OK]

Hint: Discriminant property must be named exactly 'type' [OK]

Common Mistakes:

• Using different property names like 'status' or 'state'
• Mixing properties between states incorrectly
• Adding data to loading state incorrectly

3.

Given the following TypeScript code, what will be the output?

type State =
  | { type: 'idle' }
  | { type: 'loading' }
  | { type: 'success'; data: string };

function getMessage(state: State): string {
  switch (state.type) {
    case 'idle':
      return 'Waiting...';
    case 'loading':
      return 'Loading...';
    case 'success':
      return `Data: ${state.data}`;
  }
}

console.log(getMessage({ type: 'success', data: 'Hello' }));

medium

A. Error: Property 'data' does not exist

B. Loading...

C. Waiting...

D. Data: Hello

Solution

1. Step 1: Analyze the input state

   The input state is { type: 'success', data: 'Hello' }, matching the 'success' case.

2. Step 2: Follow the switch case logic

   For 'success', the function returns `Data: ${state.data}`, which becomes 'Data: Hello'.

3. Final Answer:

   Data: Hello -> Option D

4. Quick Check:

   State 'success' returns data string [OK]

Hint: Match type in switch to get correct output [OK]

Common Mistakes:

• Confusing 'success' with 'loading' or 'idle'
• Expecting an error accessing 'data' without checking type
• Ignoring template string interpolation

4.

Identify the error in this discriminated union state machine code:

type State =
  | { type: 'start' }
  | { type: 'end'; result: number };

function process(state: State) {
  if (state.type === 'start') {
    console.log('Starting');
  } else {
    console.log('Result is ' + state.result);
  }
}

process({ type: 'start' });

medium

A. No error; code runs and prints 'Starting'.

B. Error: Accessing 'result' on 'start' state causes runtime error.

C. Error: Missing return type annotation on function.

D. Error: 'type' property should be optional.

Solution

1. Step 1: Check the state type conditions

   The function checks if state.type === 'start' and logs 'Starting'.

2. Step 2: Confirm safe access to 'result'

   In the else branch, state.result is accessed only if state.type !== 'start', so it must be 'end' with result.

3. Final Answer:

   No error; code runs and prints 'Starting'. -> Option A

4. Quick Check:

   Type narrowing prevents invalid property access [OK]

Hint: Check type guards before accessing state-specific properties [OK]

Common Mistakes:

• Assuming accessing 'result' on 'start' state causes error
• Expecting return type annotation is mandatory
• Thinking 'type' property should be optional

5.

You want to extend a discriminated union state machine to handle an error state with an optional message. Which is the best way to define this state to keep type safety and allow missing message?

type State =
  | { type: 'idle' }
  | { type: 'loading' }
  | { type: 'success'; data: string }
  | ???

hard

A. { type: 'error'; message?: string }

B. { type: 'error'; message: string | undefined }

C. { type: 'error'; message: string }

D. { type: 'error'; message: string | null }

Solution

1. Step 1: Understand optional properties in discriminated unions

   Using message?: string means the property can be missing or a string, which fits optional message.

2. Step 2: Compare other options

   { type: 'error'; message: string | undefined } requires the property but allows undefined explicitly, less idiomatic. { type: 'error'; message: string } requires message always. { type: 'error'; message: string | null } requires message but allows null, which is less type safe.

3. Final Answer:

   { type: 'error'; message?: string } -> Option A

4. Quick Check:

   Optional property syntax keeps type safety and flexibility [OK]

Hint: Use '?' to mark optional properties in state types [OK]

Common Mistakes:

• Making message required when it should be optional
• Using union with undefined instead of optional property
• Allowing null which can cause runtime issues