Skip to main content

​
Why Chain Operations?

Chaining allows you to compose multiple fallible operations without nested conditionals or try-catch blocks: 
// Without chaining - nested and hard to read
const result1 = parseInput(input)
if (result1.isOk()) {
  const result2 = validate(result1.value)
  if (result2.isOk()) {
    const result3 = save(result2.value)
    if (result3.isOk()) {
      return result3.value
    }
  }
}

// With chaining - linear and readable
const result = parseInput(input)
  .andThen(validate)
  .andThen(save)

​
Using andThen for Sequential Operations

​
Basic andThen Pattern

andThen() is used when your callback returns another Result: 
import { Result, ok, err } from 'neverthrow'

const square = (n: number): Result<number, string> => ok(n ** 2)

const result = ok(2)
  .andThen(square)  // Ok(4)
  .andThen(square)  // Ok(16)

result._unsafeUnwrap() // 16
If any step fails, the chain short-circuits: 
const divide = (n: number): Result<number, string> => {
  if (n === 0) return err('Division by zero')
  return ok(10 / n)
}

const result = ok(2)
  .andThen(square)   // Ok(4)
  .andThen(divide)   // Err('Division by zero')
  .andThen(square)   // Skipped

result._unsafeUnwrapErr() // 'Division by zero'

​
Real-World Example: User Registration

1

Define validation functions

Each function returns a Result to handle validation errors:
import { Result, ok, err } from 'neverthrow'

interface RawUser {
  email: string
  password: string
}

interface ValidUser extends RawUser {
  id: string
}

function validateEmail(user: RawUser): Result<RawUser, string> {
  if (!user.email.includes('@')) {
    return err('Invalid email format')
  }
  return ok(user)
}

function validatePassword(user: RawUser): Result<RawUser, string> {
  if (user.password.length < 8) {
    return err('Password must be at least 8 characters')
  }
  return ok(user)
}

function checkEmailUnique(user: RawUser): Result<RawUser, string> {
  // Simulate database check
  const existingUsers = ['[email protected]']
  if (existingUsers.includes(user.email)) {
    return err('Email already registered')
  }
  return ok(user)
}

function createUser(user: RawUser): Result<ValidUser, string> {
  // Simulate user creation
  return ok({
    ...user,
    id: Math.random().toString(36)
  })
}
2

Chain the operations

Compose all validation and creation steps:
function registerUser(data: RawUser): Result<ValidUser, string> {
  return ok(data)
    .andThen(validateEmail)
    .andThen(validatePassword)
    .andThen(checkEmailUnique)
    .andThen(createUser)
}

// Usage
const result = registerUser({
  email: '[email protected]',
  password: 'secure123'
})

result.match(
  (user) => console.log('User created:', user.id),
  (error) => console.error('Registration failed:', error)
)
3

Handle the result

The chain stops at the first error, providing clear feedback:
// Valid input - all steps succeed
registerUser({
  email: '[email protected]',
  password: 'securepassword'
}) // Ok(ValidUser)

// Invalid email - stops at first validation
registerUser({
  email: 'invalid',
  password: 'securepassword'
}) // Err('Invalid email format')

// Weak password - stops at second validation
registerUser({
  email: '[email protected]',
  password: 'weak'
}) // Err('Password must be at least 8 characters')

​
Flattening Nested Results

andThen automatically flattens nested Results: 
// Without andThen - nested Result
const nested = ok(ok(1234))
// Type: Ok<Ok<number, E>, E>

// With andThen - flattened
const flattened = nested.andThen(innerResult => innerResult)
// Type: Ok<number, E>
flattened._unsafeUnwrap() // 1234

​
map vs andThen

Understand when to use each method:
When your callback returns a regular value (not a Result):
const result = ok(5)
  .map(n => n * 2)        // Returns number
  .map(n => n.toString()) // Returns string

// Type: Result<string, E>

​
Error Context with mapErr

Add context to errors as they flow through the chain: 
function parseJSON(input: string): Result<unknown, string> {
  try {
    return ok(JSON.parse(input))
  } catch {
    return err('JSON_PARSE_ERROR')
  }
}

interface Config {
  apiKey: string
  timeout: number
}

function validateConfig(data: unknown): Result<Config, string> {
  if (typeof data !== 'object' || data === null) {
    return err('CONFIG_INVALID_TYPE')
  }
  const obj = data as any
  if (!obj.apiKey || !obj.timeout) {
    return err('CONFIG_MISSING_FIELDS')
  }
  return ok(obj as Config)
}

// Chain with error context
function loadConfig(jsonString: string): Result<Config, string> {
  return parseJSON(jsonString)
    .mapErr(e => `Failed to parse JSON: ${e}`)
    .andThen(validateConfig)
    .mapErr(e => `Configuration error: ${e}`)
}

const result = loadConfig('invalid json')
result._unsafeUnwrapErr() // "Failed to parse JSON: JSON_PARSE_ERROR"

​
Distinct Error Types

Starting from v4.1.0, andThen supports returning different error types: 
enum NetworkError {
  Timeout = 'TIMEOUT',
  NotFound = 'NOT_FOUND'
}

enum DatabaseError {
  ConnectionFailed = 'CONNECTION_FAILED',
  QueryFailed = 'QUERY_FAILED'
}

function fetchUser(id: string): Result<string, NetworkError> {
  // Simulate network request
  if (id === 'invalid') {
    return err(NetworkError.NotFound)
  }
  return ok(id)
}

function saveToCache(userId: string): Result<void, DatabaseError> {
  // Simulate database operation
  return ok(undefined)
}

// Error types are combined automatically
const result = fetchUser('123')
  .andThen(saveToCache)

// Type: Result<void, NetworkError | DatabaseError>

​
Advanced Pattern: Railway-Oriented Programming

Build complex pipelines that handle success and failure tracks: 
interface Order {
  id: string
  items: string[]
  total: number
}

interface ProcessedOrder extends Order {
  confirmed: boolean
  trackingNumber?: string
}

function validateOrder(order: Order): Result<Order, string> {
  if (order.items.length === 0) {
    return err('Order must contain at least one item')
  }
  if (order.total <= 0) {
    return err('Order total must be positive')
  }
  return ok(order)
}

function applyDiscount(order: Order): Result<Order, string> {
  // Apply 10% discount for orders over $100
  const discounted = order.total > 100 
    ? { ...order, total: order.total * 0.9 }
    : order
  return ok(discounted)
}

function confirmOrder(order: Order): Result<ProcessedOrder, string> {
  return ok({
    ...order,
    confirmed: true,
    trackingNumber: 'TRK-' + Math.random().toString(36).slice(2, 9)
  })
}

function processOrder(order: Order): Result<ProcessedOrder, string> {
  return ok(order)
    .andThen(validateOrder)
    .andThen(applyDiscount)
    .map(order => ({ ...order, total: Math.round(order.total * 100) / 100 }))
    .andThen(confirmOrder)
}

// Usage
const order: Order = {
  id: 'ORD-001',
  items: ['item1', 'item2'],
  total: 150
}

processOrder(order).match(
  (processed) => {
    console.log('Order confirmed!')
    console.log('Final total:', processed.total) // 135 (with discount)
    console.log('Tracking:', processed.trackingNumber)
  },
  (error) => console.error('Order failed:', error)
)

​
Next Steps

Async Operations

Learn how to chain asynchronous operations with ResultAsync

Combining Results

Work with multiple Results at once using combine

Build docs developers (and LLMs) love

Get started for freeTalk to us