Phantom types in Typescript

Introduction

Phantom types help us add extra information to types without changing the actual data. This keeps our code safer and clearer.

When you want to mark data with extra meaning without adding runtime cost.

To prevent mixing different kinds of similar data, like user IDs and product IDs.

When you want the compiler to catch mistakes about how data is used.

To add extra checks in your code without changing the data shape.

When you want to make your code easier to understand by showing intent in types.

Syntax

Typescript

type Phantom<Type, PhantomType> = Type & { __phantom?: PhantomType }

Phantom types use a type parameter that does not affect the runtime value.

The extra property (like __phantom) is optional and never used at runtime.

Examples

This example creates a UserId type that is really a string but marked specially to avoid mixing with other strings.

Typescript

type UserId = string & { __brand: 'UserId' };

function getUser(id: UserId) {
  console.log(`User ID is ${id}`);
}

This shows how phantom types prevent mixing IDs that look the same but mean different things.

Typescript

type ProductId = string & { __brand: 'ProductId' };

const userId: UserId = 'abc123' as UserId;
const productId: ProductId = 'abc123' as ProductId;

// getUser(productId); // Error: Argument of type 'ProductId' is not assignable to parameter of type 'UserId'.

Phantom types can help catch errors like adding meters and seconds, which don't make sense together.

Typescript

type Phantom<T, PhantomType> = T & { __phantom?: PhantomType };

type Meter = Phantom<number, 'Meter'>;
type Second = Phantom<number, 'Second'>;

const distance = 5 as Meter;
const time = 10 as Second;

// const wrong = distance + time; // Error: Operator '+' cannot be applied to types 'Meter' and 'Second'.

Sample Program

This program defines phantom types for user and product IDs to avoid mixing them. It shows how the compiler helps catch mistakes.

Typescript

type Phantom<T, PhantomType> = T & { __phantom?: PhantomType };

type UserId = Phantom<string, 'UserId'>;
type ProductId = Phantom<string, 'ProductId'>;

function getUser(id: UserId) {
  console.log(`User ID is ${id}`);
}

const userId = 'user-123' as UserId;
const productId = 'prod-456' as ProductId;

getUser(userId); // Works fine
// getUser(productId); // Uncommenting this line causes a compile-time error

Important Notes

Phantom types do not exist at runtime, so they add no extra cost.

Use as to tell TypeScript when you want to treat a value as a phantom type.

Phantom types help catch bugs early by making types more specific.

Summary

Phantom types add extra meaning to types without changing data.

They help prevent mixing similar but different data types.

They improve code safety and clarity with no runtime cost.

Practice

1. What is the main purpose of using phantom types in TypeScript?

easy

A. To improve performance by removing types

B. To create new objects at runtime

C. To add extra type information without affecting runtime data

D. To allow dynamic typing in TypeScript

Solution

1. Step 1: Understand phantom types concept

   Phantom types add extra type information that exists only at compile time and does not affect the actual data or runtime behavior.

2. Step 2: Compare options with concept

   Only To add extra type information without affecting runtime data correctly states that phantom types add extra type info without changing runtime data. Other options describe unrelated concepts.

3. Final Answer:

   To add extra type information without affecting runtime data -> Option C

4. Quick Check:

   Phantom types = extra compile-time info [OK]

Hint: Phantom types add info only at compile time, no runtime change [OK]

Common Mistakes:

• Thinking phantom types create new objects
• Confusing phantom types with runtime data changes
• Believing phantom types improve runtime performance

2. Which of the following is the correct way to declare a phantom type in TypeScript?

easy

A. type UserId = string & { readonly __brand: unique symbol }

B. type UserId = string | number

C. type UserId = string[]

D. type UserId = number

Solution

1. Step 1: Identify phantom type syntax

   Phantom types use intersection with a unique symbol or branded property to add phantom info without changing the base type.

2. Step 2: Check each option

   type UserId = string & { readonly __brand: unique symbol } uses intersection with a unique symbol brand, which is the correct phantom type pattern. Others are normal types or unions.

3. Final Answer:

   type UserId = string & { readonly __brand: unique symbol } -> Option A

4. Quick Check:

   Phantom type = base type & unique brand [OK]

Hint: Phantom types use intersection with unique symbol branding [OK]

Common Mistakes:

• Using union types instead of intersection for phantom types
• Declaring phantom types as arrays or simple types
• Missing the unique symbol branding

3. What will be the result of this TypeScript code?

type Meter = number & { readonly __unit: 'meter' };
function addDistance(a: Meter, b: Meter): Meter {
  return (a + b) as Meter;
}
const dist1 = 5 as Meter;
const dist2 = 10 as Meter;
const total = addDistance(dist1, dist2);
console.log(total);

medium

A. NaN

B. 15

C. Type error at addDistance call

D. undefined

Solution

1. Step 1: Understand phantom type behavior at runtime

   Phantom types exist only at compile time. At runtime, they are just numbers. So adding two numbers works normally.

2. Step 2: Analyze code output

   dist1 and dist2 are numbers with phantom type. Adding them returns 15, which is logged.

3. Final Answer:

   15 -> Option B

4. Quick Check:

   Phantom types don't affect runtime value [OK]

Hint: Phantom types vanish at runtime; values behave normally [OK]

Common Mistakes:

• Expecting type errors at runtime
• Thinking phantom types change runtime values
• Confusing phantom types with runtime wrappers

4. Identify the error in this phantom type usage:

type UserId = string & { readonly __brand: unique symbol };
function getUser(id: UserId) {
  return `User: ${id}`;
}
const id = 'abc123';
getUser(id);

medium

A. Function getUser should return a number, not string

B. No error, code runs fine

C. Unique symbol cannot be used in type intersections

D. Passing a plain string instead of branded UserId causes a type error

Solution

1. Step 1: Check argument type for getUser

   getUser expects a UserId (string with phantom brand). Passing plain string 'abc123' lacks the brand, causing a type error.

2. Step 2: Validate other options

   Return type is correct, unique symbol usage is valid, and code does not run fine due to type mismatch.

3. Final Answer:

   Passing a plain string instead of branded UserId causes a type error -> Option D

4. Quick Check:

   Phantom type mismatch = type error [OK]

Hint: Phantom types require branded values, not plain base types [OK]

Common Mistakes:

• Ignoring phantom brand when passing values
• Thinking unique symbol is invalid in types
• Assuming no error when types differ

5. How can phantom types help prevent mixing different ID types in a TypeScript project?
Consider these types:

type UserId = string & { readonly __brand: 'UserId' };
type ProductId = string & { readonly __brand: 'ProductId' };

What is the main benefit of this approach?

hard

A. It prevents accidentally using a ProductId where a UserId is expected at compile time

B. It improves runtime performance by optimizing string storage

C. It automatically converts between UserId and ProductId

D. It allows mixing IDs freely without errors

Solution

1. Step 1: Understand phantom branding for IDs

   Phantom types add unique brands to similar base types (string) to distinguish them at compile time.

2. Step 2: Identify benefit in type safety

   This branding prevents passing a ProductId where a UserId is expected, catching errors before running code.

3. Final Answer:

   It prevents accidentally using a ProductId where a UserId is expected at compile time -> Option A

4. Quick Check:

   Phantom types enforce distinct types safely [OK]

Hint: Phantom types catch ID mix-ups at compile time [OK]

Common Mistakes:

• Thinking phantom types affect runtime performance
• Assuming automatic conversion between branded types
• Believing phantom types allow mixing without errors