createAsyncThunk
Overview
A function that accepts a Redux action type string and a callback function that should return a promise. It generates promise lifecycle action types based on the action type prefix that you pass in, and returns a thunk action creator that will run the promise callback and dispatch the lifecycle actions based on the returned promise.
This abstracts the standard recommended approach for handling async request lifecycles.
It does not generate any reducer functions, since it does not know what data you're fetching, how you want to track loading state, or how the data you return needs to be processed. You should write your own reducer logic that handles these actions, with whatever loading state and processing logic is appropriate for your own app.
tip
Redux Toolkit's RTK Query data fetching API is a purpose built data fetching and caching solution for Redux apps, and can eliminate the need to write any thunks or reducers to manage data fetching. We encourage you to try it out and see if it can help simplify the data fetching code in your own apps!
Sample usage:
import { createAsyncThunk, createSlice } from '@reduxjs/toolkit'
import { userAPI } from './userAPI'
// First, create the thunk
const fetchUserById = createAsyncThunk(
'users/fetchByIdStatus',
async (userId: number, thunkAPI) => {
const response = await userAPI.fetchById(userId)
return response.data
}
)
interface UsersState {
entities: []
loading: 'idle' | 'pending' | 'succeeded' | 'failed'
}
const initialState = {
entities: [],
loading: 'idle',
} as UsersState
// Then, handle actions in your reducers:
const usersSlice = createSlice({
name: 'users',
initialState,
reducers: {
// standard reducer logic, with auto-generated action types per reducer
},
extraReducers: (builder) => {
// Add reducers for additional action types here, and handle loading state as needed
builder.addCase(fetchUserById.fulfilled, (state, action) => {
// Add user to the state array
state.entities.push(action.payload)
})
},
})
// Later, dispatch the thunk as needed in the app
dispatch(fetchUserById(123))
Parameters
createAsyncThunk
accepts three parameters: a string action type
value, a payloadCreator
callback, and an options
object.
type
A string that will be used to generate additional Redux action type constants, representing the lifecycle of an async request:
For example, a type
argument of 'users/requestStatus'
will generate these action types:
pending
:'users/requestStatus/pending'
fulfilled
:'users/requestStatus/fulfilled'
rejected
:'users/requestStatus/rejected'
payloadCreator
A callback function that should return a promise containing the result of some asynchronous logic. It may also return a value synchronously. If there is an error, it should either return a rejected promise containing an Error
instance or a plain value such as a descriptive error message or otherwise a resolved promise with a RejectWithValue
argument as returned by the thunkAPI.rejectWithValue
function.
The payloadCreator
function can contain whatever logic you need to calculate an appropriate result. This could include a standard AJAX data fetch request, multiple AJAX calls with the results combined into a final value, interactions with React Native AsyncStorage
, and so on.
The payloadCreator
function will be called with two arguments:
arg
: a single value, containing the first parameter that was passed to the thunk action creator when it was dispatched. This is useful for passing in values like item IDs that may be needed as part of the request. If you need to pass in multiple values, pass them together in an object when you dispatch the thunk, likedispatch(fetchUsers({status: 'active', sortBy: 'name'}))
.thunkAPI
: an object containing all of the parameters that are normally passed to a Redux thunk function, as well as additional options:dispatch
: the Redux storedispatch
methodgetState
: the Redux storegetState
methodextra
: the "extra argument" given to the thunk middleware on setup, if availablerequestId
: a unique string ID value that was automatically generated to identify this request sequencesignal
: anAbortController.signal
object that may be used to see if another part of the app logic has marked this request as needing cancelation.rejectWithValue(value, [meta])
: rejectWithValue is a utility function that you canreturn
(orthrow
) in your action creator to return a rejected response with a defined payload and meta. It will pass whatever value you give it and return it in the payload of the rejected action. If you also pass in ameta
, it will be merged with the existingrejectedAction.meta
.fulfillWithValue(value, meta)
: fulfillWithValue is a utility function that you canreturn
in your action creator tofulfill
with a value while having the ability of adding tofulfilledAction.meta
.
The logic in the payloadCreator
function may use any of these values as needed to calculate the result.
Options
An object with the following optional fields:
condition(arg, { getState, extra } ): boolean | Promise<boolean>
: a callback that can be used to skip execution of the payload creator and all action dispatches, if desired. See Canceling Before Execution for a complete description.dispatchConditionRejection
: ifcondition()
returnsfalse
, the default behavior is that no actions will be dispatched at all. If you still want a "rejected" action to be dispatched when the thunk was canceled, set this flag totrue
.idGenerator(arg): string
: a function to use when generating therequestId
for the request sequence. Defaults to use nanoid, but you can implement your own ID generation logic.serializeError(error: unknown) => any
to replace the internalminiSerializeError
method with your own serialization logic.getPendingMeta({ arg, requestId }, { getState, extra }): any
: a function to create an object that will be merged into thependingAction.meta
field.
Return Value
createAsyncThunk
returns a standard Redux thunk action creator. The thunk action creator function will have plain action creators for the pending
, fulfilled
, and rejected
cases attached as nested fields.
Using the fetchUserById
example above, createAsyncThunk
will generate four functions:
fetchUserById
, the thunk action creator that kicks off the async payload callback you wrotefetchUserById.pending
, an action creator that dispatches an'users/fetchByIdStatus/pending'
actionfetchUserById.fulfilled
, an action creator that dispatches an'users/fetchByIdStatus/fulfilled'
actionfetchUserById.rejected
, an action creator that dispatches an'users/fetchByIdStatus/rejected'
action
When dispatched, the thunk will:
- dispatch the
pending
action - call the
payloadCreator
callback and wait for the returned promise to settle - when the promise settles:
- if the promise resolved successfully, dispatch the
fulfilled
action with the promise value asaction.payload
- if the promise resolved with a
rejectWithValue(value)
return value, dispatch therejected
action with the value passed intoaction.payload
and 'Rejected' asaction.error.message
- if the promise failed and was not handled with
rejectWithValue
, dispatch therejected
action with a serialized version of the error value asaction.error
- if the promise resolved successfully, dispatch the
- Return a fulfilled promise containing the final dispatched action (either the
fulfilled
orrejected
action object)
Promise Lifecycle Actions
createAsyncThunk
will generate three Redux action creators using createAction
: pending
, fulfilled
, and rejected
. Each lifecycle action creator will be attached to the returned thunk action creator so that your reducer logic can reference the action types and respond to the actions when dispatched. Each action object will contain the current unique requestId
and arg
values under action.meta
.
The action creators will have these signatures:
interface SerializedError {
name?: string
message?: string
code?: string
stack?: string
}
interface PendingAction<ThunkArg> {
type: string
payload: undefined
meta: {
requestId: string
arg: ThunkArg
}
}
interface FulfilledAction<ThunkArg, PromiseResult> {
type: string
payload: PromiseResult
meta: {
requestId: string
arg: ThunkArg
}
}
interface RejectedAction<ThunkArg> {
type: string
payload: undefined
error: SerializedError | any
meta: {
requestId: string
arg: ThunkArg
aborted: boolean
condition: boolean
}
}
interface RejectedWithValueAction<ThunkArg, RejectedValue> {
type: string
payload: RejectedValue
error: { message: 'Rejected' }
meta: {
requestId: string
arg: ThunkArg
aborted: boolean
}
}
type Pending = <ThunkArg>(
requestId: string,
arg: ThunkArg
) => PendingAction<ThunkArg>
type Fulfilled = <ThunkArg, PromiseResult>(
payload: PromiseResult,
requestId: string,
arg: ThunkArg
) => FulfilledAction<ThunkArg, PromiseResult>
type Rejected = <ThunkArg>(
requestId: string,
arg: ThunkArg
) => RejectedAction<ThunkArg>
type RejectedWithValue = <ThunkArg, RejectedValue>(
requestId: string,
arg: ThunkArg
) => RejectedWithValueAction<ThunkArg, RejectedValue>
To handle these actions in your reducers, reference the action creators in createReducer
or createSlice
using either the object key notation or the "builder callback" notation. (Note that if you use TypeScript, you should use the "builder callback" notation to ensure the types are inferred correctly):
const reducer1 = createReducer(initialState, {
[fetchUserById.fulfilled]: (state, action) => {},
})
const reducer2 = createReducer(initialState, (builder) => {
builder.addCase(fetchUserById.fulfilled, (state, action) => {})
})
const reducer3 = createSlice({
name: 'users',
initialState,
reducers: {},
extraReducers: {
[fetchUserById.fulfilled]: (state, action) => {},
},
})
const reducer4 = createSlice({
name: 'users',
initialState,
reducers: {},
extraReducers: (builder) => {
builder.addCase(fetchUserById.fulfilled, (state, action) => {})
},
})
Handling Thunk Results
Unwrapping Result Actions
Thunks may return a value when dispatched. A common use case is to return a promise from the thunk, dispatch the thunk from a component, and then wait for the promise to resolve before doing additional work:
const onClick = () => {
dispatch(fetchUserById(userId)).then(() => {
// do additional work
})
}
The thunks generated by createAsyncThunk
will always return a resolved promise with either the fulfilled
action object or rejected
action object inside, as appropriate.
The calling logic may wish to treat these actions as if they were the original promise contents. The promise returned by the dispatched thunk has an unwrap
property which can be called to extract the payload
of a fulfilled
action or to throw either the error
or, if available, payload
created by rejectWithValue
from a rejected
action:
// in the component
const onClick = () => {
dispatch(fetchUserById(userId))
.unwrap()
.then((originalPromiseResult) => {
// handle result here
})
.catch((rejectedValueOrSerializedError) => {
// handle error here
})
}
Or with async/await syntax:
// in the component
const onClick = async () => {
try {
const originalPromiseResult = await dispatch(fetchUserById(userId)).unwrap()
// handle result here
} catch (rejectedValueOrSerializedError) {
// handle error here
}
}
Using the attached .unwrap()
property is preferred in most cases, however Redux Toolkit also exports an unwrapResult
function that can be used for a similar purpose:
import { unwrapResult } from '@reduxjs/toolkit'
// in the component
const onClick = () => {
dispatch(fetchUserById(userId))
.then(unwrapResult)
.then((originalPromiseResult) => {
// handle result here
})
.catch((rejectedValueOrSerializedError) => {
// handle result here
})
}
Or with async/await syntax:
import { unwrapResult } from '@reduxjs/toolkit'
// in the component
const onClick = async () => {
try {
const resultAction = await dispatch(fetchUserById(userId))
const originalPromiseResult = unwrapResult(resultAction)
// handle result here
} catch (rejectedValueOrSerializedError) {
// handle error here
}
}
Checking Errors After Dispatching
Note that this means a failed request or error in a thunk will never return a rejected promise. We assume that any failure is more of a handled error than an unhandled exception at this point. This is due to the fact that we want to prevent uncaught promise rejections for those who do not use the result of dispatch
.
If your component needs to know if the request failed, use .unwrap
or unwrapResult
and handle the re-thrown error accordingly.
Handling Thunk Errors
When your payloadCreator
returns a rejected promise (such as a thrown error in an async
function), the thunk will dispatch a rejected
action containing an automatically-serialized version of the error as action.error
. However, to ensure serializability, everything that does not match the SerializedError
interface will have been removed from it:
export interface SerializedError {
name?: string
message?: string
stack?: string
code?: string
}
If you need to customize the contents of the rejected
action, you should catch any errors yourself, and then return a new value using the thunkAPI.rejectWithValue
utility. Doing return rejectWithValue(errorPayload)
will cause the rejected
action to use that value as action.payload
.
The rejectWithValue
approach should also be used if your API response "succeeds", but contains some kind of additional error details that the reducer should know about. This is particularly common when expecting field-level validation errors from an API.
const updateUser = createAsyncThunk(
'users/update',
async (userData, { rejectWithValue }) => {
const { id, ...fields } = userData
try {
const response = await userAPI.updateById(id, fields)
return response.data.user
} catch (err) {
// Use `err.response.data` as `action.payload` for a `rejected` action,
// by explicitly returning it using the `rejectWithValue()` utility
return rejectWithValue(err.response.data)
}
}
)
Cancellation
Canceling Before Execution
If you need to cancel a thunk before the payload creator is called, you may provide a condition
callback as an option after the payload creator. The callback will receive the thunk argument and an object with {getState, extra}
as parameters, and use those to decide whether to continue or not. If the execution should be canceled, the condition
callback should return a literal false
value or a promise that should resolve to false
. If a promise is returned, the thunk waits for it to get fulfilled before dispatching the pending
action, otherwise it proceeds with dispatching synchronously.
const fetchUserById = createAsyncThunk(
'users/fetchByIdStatus',
async (userId: number, thunkAPI) => {
const response = await userAPI.fetchById(userId)
return response.data
},
{
condition: (userId, { getState, extra }) => {
const { users } = getState()
const fetchStatus = users.requests[userId]
if (fetchStatus === 'fulfilled' || fetchStatus === 'loading') {
// Already fetched or in progress, don't need to re-fetch
return false
}
},
}
)
If condition()
returns false
, the default behavior is that no actions will be dispatched at all. If you still want a "rejected" action to be dispatched when the thunk was canceled, pass in {condition, dispatchConditionRejection: true}
.
Canceling While Running
If you want to cancel your running thunk before it has finished, you can use the abort
method of the promise returned by dispatch(fetchUserById(userId))
.
A real-life example of that would look like this:
// file: store.ts noEmit
import { configureStore } from '@reduxjs/toolkit'
import type { Reducer } from '@reduxjs/toolkit'
import { useDispatch } from 'react-redux'
declare const reducer: Reducer<{}>
const store = configureStore({ reducer })
export const useAppDispatch = () => useDispatch<typeof store.dispatch>()
// file: slice.ts noEmit
import { createAsyncThunk } from '@reduxjs/toolkit'
export const fetchUserById = createAsyncThunk(
'fetchUserById',
(userId: string) => {
/* ... */
}
)
// file: MyComponent.ts
import { fetchUserById } from './slice'
import { useAppDispatch } from './store'
import React from 'react'
function MyComponent(props: { userId: string }) {
const dispatch = useAppDispatch()
React.useEffect(() => {
// Dispatching the thunk returns a promise
const promise = dispatch(fetchUserById(props.userId))
return () => {
// `createAsyncThunk` attaches an `abort()` method to the promise
promise.abort()
}
}, [props.userId])
}
After a thunk has been cancelled this way, it will dispatch (and return) a "thunkName/rejected"
action with an AbortError
on the error
property. The thunk will not dispatch any further actions.
Additionally, your payloadCreator
can use the AbortSignal
it is passed via thunkAPI.signal
to actually cancel a costly asynchronous action.
The fetch
api of modern browsers already comes with support for an AbortSignal
:
import { createAsyncThunk } from '@reduxjs/toolkit'
const fetchUserById = createAsyncThunk(
'users/fetchById',
async (userId: string, thunkAPI) => {
const response = await fetch(`https://reqres.in/api/users/${userId}`, {
signal: thunkAPI.signal,
})
return await response.json()
}
)
Checking Cancellation Status
Reading the Signal Value
You can use the signal.aborted
property to regularly check if the thunk has been aborted and in that case stop costly long-running work:
import { createAsyncThunk } from '@reduxjs/toolkit'
const readStream = createAsyncThunk(
'readStream',
async (stream: ReadableStream, { signal }) => {
const reader = stream.getReader()
let done = false
let result = ''
while (!done) {
if (signal.aborted) {
throw new Error('stop the work, this has been aborted!')
}
const read = await reader.read()
result += read.value
done = read.done
}
return result
}
)
Listening for Abort Events
You can also call signal.addEventListener('abort', callback)
to have logic inside the thunk be notified when promise.abort()
was called.
This can for example be used in conjunction with an axios CancelToken
:
import { createAsyncThunk } from '@reduxjs/toolkit'
import axios from 'axios'
const fetchUserById = createAsyncThunk(
'users/fetchById',
async (userId: string, { signal }) => {
const source = axios.CancelToken.source()
signal.addEventListener('abort', () => {
source.cancel()
})
const response = await axios.get(`https://reqres.in/api/users/${userId}`, {
cancelToken: source.token,
})
return response.data
}
)
Checking if a Promise Rejection was from an Error or Cancellation
To investigate behavior around thunk cancellation, you can inspect various properties on the meta
object of the dispatched action.
If a thunk was cancelled, the result of the promise will be a rejected
action (regardless of whether that action was actually dispatched to the store).
- If it was cancelled before execution,
meta.condition
will be true. - If it was aborted while running,
meta.aborted
will be true. - If neither of those is true, the thunk was not cancelled, it was simply rejected, either by a Promise rejection or
rejectWithValue
. - If the thunk was not rejected, both
meta.aborted
andmeta.condition
will beundefined
.
So if you wanted to test that a thunk was cancelled before executing, you can do the following:
import { createAsyncThunk, isRejected } from '@reduxjs/toolkit'
test('this thunk should always be skipped', async () => {
const thunk = createAsyncThunk(
'users/fetchById',
async () => throw new Error('This promise should never be entered'),
{
condition: () => false,
}
)
const result = await thunk()(dispatch, getState, null)
expect(result.meta.condition).toBe(true)
expect(result.meta.aborted).toBe(false)
})
Examples
- Requesting a user by ID, with loading state, and only one request at a time:
import { createAsyncThunk, createSlice } from '@reduxjs/toolkit'
import { userAPI, User } from './userAPI'
const fetchUserById = createAsyncThunk<User, string, {
state: { users: { loading: string, currentRequestId: string } }
}> (
'users/fetchByIdStatus',
async (userId: string, { getState, requestId }) => {
const { currentRequestId, loading } = getState().users
if (loading !== 'pending' || requestId !== currentRequestId) {
return
}
const response = await userAPI.fetchById(userId)
return response.data
}
)
const usersSlice = createSlice({
name: 'users',
initialState: {
entities: [],
loading: 'idle',
currentRequestId: undefined,
error: null,
},
reducers: {},
extraReducers: (builder) => {
builder
.addCase(fetchUserById.pending, (state, action) => {
if (state.loading === 'idle') {
state.loading = 'pending'
state.currentRequestId = action.meta.requestId
}
})
.addCase(fetchUserById.fulfilled, (state, action) => {
const { requestId } = action.meta
if (
state.loading === 'pending' &&
state.currentRequestId === requestId
) {
state.loading = 'idle'
state.entities.push(action.payload)
state.currentRequestId = undefined
}
})
.addCase(fetchUserById.rejected, (state, action) => {
const { requestId } = action.meta
if (
state.loading === 'pending' &&
state.currentRequestId === requestId
) {
state.loading = 'idle'
state.error = action.error
state.currentRequestId = undefined
}
})
},
})
const UsersComponent = () => {
const { entities, loading, error } = useSelector((state) => state.users)
const dispatch = useDispatch()
const fetchOneUser = async (userId) => {
try {
const user = await dispatch(fetchUserById(userId)).unwrap()
showToast('success', `Fetched ${user.name}`)
} catch (err) {
showToast('error', `Fetch failed: ${err.message}`)
}
}
// render UI here
}
Using rejectWithValue to access a custom rejected payload in a component
Note: this is a contrived example assuming our userAPI only ever throws validation-specific errors
// file: store.ts noEmit
import { configureStore } from '@reduxjs/toolkit'
import type { Reducer } from '@reduxjs/toolkit'
import { useDispatch } from 'react-redux'
import usersReducer from './user/slice'
const store = configureStore({ reducer: { users: usersReducer } })
export const useAppDispatch = () => useDispatch<typeof store.dispatch>()
export type RootState = ReturnType<typeof store.getState>
// file: user/userAPI.ts noEmit
export declare const userAPI: {
updateById<Response>(id: string, fields: {}): { data: Response }
}
// file: user/slice.ts
import { createAsyncThunk, createSlice } from '@reduxjs/toolkit'
import { userAPI } from './userAPI'
import type { AxiosError } from 'axios'
// Sample types that will be used
export interface User {
id: string
first_name: string
last_name: string
email: string
}
interface ValidationErrors {
errorMessage: string
field_errors: Record<string, string>
}
interface UpdateUserResponse {
user: User
success: boolean
}
export const updateUser = createAsyncThunk<
User,
{ id: string } & Partial<User>,
{
rejectValue: ValidationErrors
}
>('users/update', async (userData, { rejectWithValue }) => {
try {
const { id, ...fields } = userData
const response = await userAPI.updateById<UpdateUserResponse>(id, fields)
return response.data.user
} catch (err) {
let error: AxiosError<ValidationErrors> = err // cast the error for access
if (!error.response) {
throw err
}
// We got validation errors, let's return those so we can reference in our component and set form errors
return rejectWithValue(error.response.data)
}
})
interface UsersState {
error: string | null | undefined
entities: Record<string, User>
}
const initialState = {
entities: {},
error: null,
} as UsersState
const usersSlice = createSlice({
name: 'users',
initialState,
reducers: {},
extraReducers: (builder) => {
// The `builder` callback form is used here because it provides correctly typed reducers from the action creators
builder.addCase(updateUser.fulfilled, (state, { payload }) => {
state.entities[payload.id] = payload
})
builder.addCase(updateUser.rejected, (state, action) => {
if (action.payload) {
// Being that we passed in ValidationErrors to rejectType in `createAsyncThunk`, the payload will be available here.
state.error = action.payload.errorMessage
} else {
state.error = action.error.message
}
})
},
})
export default usersSlice.reducer
// file: externalModules.d.ts noEmit
declare module 'some-toast-library' {
export function showToast(type: string, message: string)
}
// file: user/UsersComponent.ts
import React from 'react'
import { useAppDispatch } from '../store'
import type { RootState } from '../store'
import { useSelector } from 'react-redux'
import { updateUser } from './slice'
import type { User } from './slice'
import type { FormikHelpers } from 'formik'
import { showToast } from 'some-toast-library'
interface FormValues extends Omit<User, 'id'> {}
const UsersComponent = (props: { id: string }) => {
const { entities, error } = useSelector((state: RootState) => state.users)
const dispatch = useAppDispatch()
// This is an example of an onSubmit handler using Formik meant to demonstrate accessing the payload of the rejected action
const handleUpdateUser = async (
values: FormValues,
formikHelpers: FormikHelpers<FormValues>
) => {
const resultAction = await dispatch(updateUser({ id: props.id, ...values }))
if (updateUser.fulfilled.match(resultAction)) {
// user will have a type signature of User as we passed that as the Returned parameter in createAsyncThunk
const user = resultAction.payload
showToast('success', `Updated ${user.first_name} ${user.last_name}`)
} else {
if (resultAction.payload) {
// Being that we passed in ValidationErrors to rejectType in `createAsyncThunk`, those types will be available here.
formikHelpers.setErrors(resultAction.payload.field_errors)
} else {
showToast('error', `Update failed: ${resultAction.error}`)
}
}
}
// render UI here
}