I am learning a react/redux and have an application with two main pieces of state:
An array of items
An object that contains user-specified filters for those items
I have three functions/actions, createFilter, updateFilter, and deleteFilter that modify the state of #2. I have an action filterItems that modifies #1 based on the state of #2. So whenever #2 changes, this action needs to be dispatched.
This is the component I am working with:
import React, { Component } from 'react'
import { connect } from 'react-redux'
import { bindActionCreators } from 'redux'
import { createFilter } from '../actions/actions'
import { updateFilter } from '../actions/actions'
import { deleteFilter } from '../actions/actions'
import { filterItems } from '../actions/actions'
class ItemList extends Component {
createFilter(input) {
this.props.createFilter(input)
this.props.filterItems()
}
updateFilter(input) {
this.props.updateFilter(input)
this.props.filterItems()
}
deleteFilter() {
this.props.deleteFilter()
this.props.filterItems()
}
...
// Render method
...
}
function mapDispatchToProps(dispatch) {
return bindActionCreators({ createFilter, updateFilter, deleteFilter, filterItems }, dispatch)
}
function mapStateToProps({ itemList }) {
return { itemList }
}
export default connect(mapStateToProps, mapDispatchToProps)(ItemList)
What I have found is that when one of the filter methods are sent, the store (state #2) is not yet updated by the time filterItems() is called.
So I need to asynchronously execute the filter functions, and once the store is updated call filterItems.
I am struggling on how to do this with react-thunk. If the first function was an ajax promise I would use .then():
export function updateFilterAndEvaluate(input) {
return (dispatch, getState) => {
updateFilter(input).then(dispatch(filterItems(getState().filters)))
}
}
But these are just functions, and don't have a .then() method. I am trying to figure out what my best course of action is for this implementation. Can I wrap Redux actions in a promise? Am I misusing Thunk? Or should I attempt a different pattern entirely?
I have an action filterItems that modifies #1 based on the state of #2.
This is, generally speaking, an anti-pattern. Since the result array can be computed from the source array and the currently active filters, you shouldn’t be keeping it in the state.
Redux actions should generally look like “events” (e.g. what happened). “Filter was created” and “filter was updated” are good actions. “Filter them now!” looks more like a command, this is usually a sign that it shouldn’t have been an action in the first place, and should be something the components do as they select the data to render.
Instead, do the filtering as part of your mapStateToProps() function when you prepare data for the components. If it gets expensive, look into using Reselect to compute derived data efficiently.
As for your specific question,
What I have found is that when one of the filter methods are sent, the store (state #2) is not yet updated by the time filterItems() is called.
This is incorrect and indicates some other problem in your code. (It’s hard to tell where because the example is incomplete). In Redux, dispatch() is synchronous (unless you have some middleware that delays or batches it which usually isn’t the case), so you don’t need to “wait” for it to finish if it just operates on the local data.
However, in any case, filterItems() is not a very good fit for an action, and I suggest you to look into filtering in mapStateToProps() as I wrote above.
Related
I am new to React, and when I was reading about the docs, I found there were two ways to implement React components, functional-based and class-based. I know before React 16.8 it's not possible to manage state in functional components, but after that there is React Hooks.
The problem is, there seems to be one restriction for React Hooks, they can only be used inside functional components. Take a server-client as an example, which needs to change an isAuthenticated state while 401 received.
//client.js
import { useUserDispatch, signOut } from "auth";
export function request(url, args) {
var dispatch = useUserDispatch();
return fetch(url, args).then(response => {
if (response.status === 401) {
logout(dispatch);
}
}
);
//auth.js
import React from "react";
var UserStateContext = React.createContext();
var UserDispatchContext = React.createContext();
function userReducer(state, action) {
...
}
function UserProvider({ children }) {
var [state, dispatch] = React.useReducer(userReducer, {
isAuthenticated: false,
});
return (
<UserStateContext.Provider value={state}>
<UserDispatchContext.Provider value={dispatch}>
{children}
</UserDispatchContext.Provider>
</UserStateContext.Provider>
);
}
function useUserState() {
return React.useContext(UserStateContext);
}
function useUserDispatch() {
return React.useContext(UserDispatchContext);
}
function signOut(dispatch) {
dispatch({});
}
export { UserProvider, useUserState, useUserDispatch, loginUser, signOut };
The client code above will produce error "Hooks can only be called inside of the body of a function component".
So maybe I have to move line var dispatch = useUserDispatch() upward to the component where request is called, and pass dispatch as props to request.
I feel this is not right, no only request is forced to care about some meaningless(to it) dispatch, but also this dispatch will spread everywhere a component needs to request.
For class-based components, this.state doesn't solve this problem either, but at least I can use mobx.
So are there some other ideal ways to solve this problem?
I came at this point too. Long story short you need to use Redux and Thunk with Async Logic, as described in detail with examples in the link below [1] if you want to do all of the stuff by hand on your own.
[1] https://redux.js.org/tutorials/essentials/part-5-async-logic
There is another solution that gives out-of-the box experience with Asynchronous API (can work with OpenAPI and GraphQL, handles request, provides caching with lifecycle, etc) wrapping stuff from [1] and its called RTK Query [2].
[2] https://redux-toolkit.js.org/rtk-query/overview
Diagram below explains [1] process visually.. but I think RTK Query [2] wraps everything in one place and could be better solution. There is a Quick Start Guide [3]. I will give it a try :-)
[3] https://redux-toolkit.js.org/tutorials/rtk-query/
Mobx and hooks are very similar in implementation. Both use a render context that is in a sense "global". React ties that render context to the component render context, but Mobx keeps that render context separate. Therefore that means that hooks have to be created within a component render lifecycle (but can sometimes be called outside that context). Mobx-react ties the Mobx render lifecycle to the react lifecycle, triggering a react re-render when observed objects change. So Mobx-react nests the react render context within the Mobx render context.
React internally keeps tracks of hooks by the number of times and order the hook is called within a component render cycle. Mobx, on the other hand, wraps any "observable" object with a proxy that lets the Mobx context know if any of its properties were referenced during a Mobx "run context" (an autorun call, essentially). Then when a property is changed, Mobx knows what "run contexts" care about that property, and re-runs those contexts. This means that anywhere you have access to an observable object you can change a property on it and Mobx will react to it.
For react state hooks, react provides a custom setter function for a state object. React then uses calls to that setter to know when it needs to re-render a component. That setter can be used anywhere, even outside a React render, but you can only create that hook inside a render call, because otherwise react has no way to tell what component to tie that hook to. Creating a hook implicitly connects it to the current render context, and that's why hooks have to be created inside render calls: hook builders have no meaning outside a render call, because they have no way to know what component they are connected to -- but once tied to a component, then they need to be available anywhere. In fact, actions like onClick or a fetch callback don't occur within a render context, although the callback is often created within that context - the action callback happens after react finishes rendering (because javascript is single threaded, so the render function must complete before anything else happens).
Hooks comes as an alternatively to class based components, you should pick up one to your project and stick to it, don't mix it up. there are some motivation for the creation of hooks, as it's better stated at docs: hook motivation.
you can create hook functions apart, but they are meant to be consumed by components. it's something like using HOC (high order component) with class based components.
const myHook = () => {
[foo, setFoo] = useState('john')
// use effect for example if you need to run something at state updates
useEffect(() => {
// do something on foo changes
}, [foo])
return [foo, setFoo] // returning state and setState you can use them by your component
}
now you have a reusable hook and you can consume at your components:
const myComponent = (props) => {
[foo, setFoo] = myHook()
const handleFoo = () => {
// some logic
setFoo(someValue)
}
return (
<div>
<span>{foo}<span>
<button onClick={handleFoo}>click</button>
</div>
)
}
obs: you should avoid declare variables as var nowadays, pick const for most, and if it's a value variable (like number) that needs update use let.
When you are creating a hooks you must refer to the Rules of Hooks
You can only call hooks from a react functions.
Don’t call Hooks from regular JavaScript functions. Instead, you can:
✅ Call Hooks from React function components.
✅ Call Hooks from custom Hooks (learn about them on this page).
If you want to create a reusable hooks then you can create a custom hooks for your functions.
You can call as many functions inside a hooks.
For example, here I'm refactoring the request function as a hook.
export function useRequest(url, args) {
var userDispatch = useUserDispatch();
const fetcher = React.useCallback(() => {
return new Promise((resolve, reject) =>
fetch(url, args)
.then((response) => {
if (response.status === 401) {
logout();
reject();
}
resolve(response);
})
.catch(reject)
);
}, [url, args]);
return [fetcher, userDispatch];
}
and then consumes it.
function App() {
const [fetch, userDispatch] = useRequest("/url", {});
React.useEffect(() => {
fetch().then((response) => {
userDispatch({ type: "USER_REQUEST", payload: response });
});
}, []);
return <div>Hello world</div>;
}
Yes, you have to use Redux or MobX to solve this problem. You have to maintain isAuthenticated state in the global state of Redux or MobX. Then make an action that could be named like, toggleAuthState and pass is to the child component and toggle the state from there.
Also you can use functional components for this case. Class based components is not mandatory to use MobX or Redux. If you maintain a HOC as a Container then you can pass the actions and states to the child.
I am showing an example of using a container as a HOC:
// Container
import React from "react"
import * as actions from "../actions"
import ChildComponent from "../components/ChildComponent"
import { connect } from "react-redux"
import { bindActionCreators } from "redux"
const Container = props => <ChildComponent { ...props } />
const mapStateToProps = state => ({ ...state })
const mapDispatchToProps = dispatch => bindActionCreators(actions, dispatch)
export default connect(mapStateToProps, mapDispatchToProps)(Container)
Then in ChildComponent you can use your states and dispatch actions whenever you need.
I stumbled upon a requirement on a section of a vanilla JS webapp that requires a single JSON "definitions" object to render. Definitions are loaded via an HTTP request at the very beginning, read, parsed and handed down to another layer of the app. The object itself never changes throughout its life cycle.
I'm now trying to model this scenario in ReactJS, using Redux + redux-thunk. I created a thunk/async action the fetches the JSON object, extracts what it needs and ends up updating the state with that -
but it does not store the object itself in the state. This seems like the right, logical approach since, as mentioned, the definitions are never modified in any way. I'd argue it's simply not state, in a strict sense.
However, by taking that decision I ended up struggling while implementing the actual React.Component. Almost every single example I've seen out there in the wild for async cases like this one:
Defines a thunk action that fires some API call.
Stores whatever they got back (or after some alterations) in a state property.
Maps that property to this.props in the Component with mapStateToProps and connect.
In my case, I don't really have a state property to bind to. So I ended up returning the definitions object in my async action and using the component's local state to get what I needed.
class ContainerComponent extends React.Component {
state = { definitions: {} };
componentDidMount() {
const { dispatch } = this.props;
dispatch(fetchDefinitions())
.then((definitions) => this.setState({ definitions }));
}
render() {
return (<PresentationalComponent definitions={this.state.definitions} />);
}
}
export default connect()(ContainerComponent);
Not saying that this.setState should be avoided, but this looks an awful lot like what I had before even introducing Redux: an API call returning a promise - only with a lot more meddling indirections.
componentDidMount() {
const { dispatch } = this.props;
fetch(`${API_URL}/definitions`)
.then((res) => res.json())
.then((definitions) => this.setState({ definitions }));
}
So, how should I go about this? Is there any particular thing I am missing here? Any pattern I should be following? Perhaps, avoiding Redux entirely for this matter?
You are right in that having a component state isn't necessarily a bad thing, but I believe you are confused on where to store that data once the API call is made.
You mention that it is not necessarily state, but I would argue otherwise. Prior to making the API call, your application does not have that data. You may have certain UX/UI indications at the start up of your application that, for example could indicate on if the data is being fetched: definitions.all.isFetching.
In your componentDidMount, dispatching the action to fetch the data is correct. Once the action is fired, and the success response is received, your reducer should save the definitions to your redux store like
import { assign, get } from 'lodash';
const all = (
state = { isFetching: false, data: [] },
action,
) => {
switch (action.type) {
case types.LIST_DEFINITIONS:
return assign({}, state, { isFetching: true });
case types.LIST_DEFINITIONS_SUCCESS:
return assign({}, state, get(action, 'result.data'), { isFetching: false });
default: return state;
}
};
Then in your component, you would connect your redux store
function mapStateToProps(state){
return {
definitions: state.definitions.all.data
};
}
export default connect(mapStateToProps, { listDefinitions })(ContainerComponent);
Also note I moved the action out in my example and am placing it into the connect with mapDispatchToProps shorthand.
I am trying to make multiple changes to the store, but not render till all changes are done. I wanted to do this with redux-thunk.
Here is my action creator:
function addProp(name, value) {
return { type:'ADD_PROP', name, value }
}
function multiGeoChanges(...changes) {
// my goal here is to make multiple changes to geo, and make sure that react doesnt update the render till the end
return async function(dispatch, getState) {
for (let change of changes) {
dispatch(change);
await promiseTimeout(2000);
}
}
}
I dispatch my async action creator like this:
store.dispatch(multiGeoChanges(addProp(1, "val1"), addProp(2, "val2"), addProp(3, "val3")));
However this is causing react to render after each dispatch. I am new to redux-thunk, I never used async middleware, but I thought it could help me here.
#Kokovin Vladislav's answer is correct. To add some additional context:
Redux will notify all subscribers after every dispatch. To cut down on re-renders, either dispatch fewer times, or use one of several approaches for "batching" dispatches and notifications. For more info, see the Redux FAQ on update events: http://redux.js.org/docs/faq/Performance.html#performance-update-events .
I also recently wrote a couple of blog posts that relate to this topic. Idiomatic Redux: Thoughts on Thunks, Sagas, Abstraction, and Reusability discusses the pros and cons of using thunks, and summarizes several ways to handle batching of dispatches. Practical Redux Part 6: Connected Lists, Forms, and Performance describes several key aspects to be aware of regarding Redux performance.
Finally, there's several other libraries that can help with batching up store change notifications. See the Store#Store Change Subscriptions section of my Redux addons catalog for a list of relevant addons. In particular, you might be interested in https://github.com/manaflair/redux-batch , which will allow you to dispatch an array of actions with only a single notification event.
There are ways to achieve the goal:
Classic way:
usually:
Actions describe the fact that something happened, but don't specify how the application's state changes in response. This is the job of reducers.
That also means that actions are not setters.
Thus, you could describe what has happened and accumulate changes, and dispatch one action
something like:
const multipleAddProp = (changedProps) =>({
type:'MULTIPLE_ADD_PROP', changedProps
});
And then react on action in reducer:
const geo=(state,action)=>{
...
switch (action.type){
case 'MULTIPLE_ADD_PROP':
// apply new props
...
}
}
Another way When rerendering is critical :
then you can consider to limit components, which could be rerendered on state change.
For example you can use shouldComponentUpdate to check whether component
should be rendered or not.
Also you could use reselect, in order to not rerender connected components
after calculating derived data...
Non standard way:
redux-batched-action
It works something like transaction.
In this example, the subscribers would be notified once:
import { batchActions } from 'redux-batched-actions';
const multiGeoChanges=(...arrayOfActions)=> dispatch => {
dispatch( batchActions(arrayOfActions) );
}
In react-redux 7.0.1+ batching is now built-in. Release notes of 7.0.1:
https://github.com/reduxjs/react-redux/releases/tag/v7.0.1
Batched Updates
React has an unstable_batchedUpdates API that it uses to group
together multiple updates from the same event loop tick. The React
team encouraged us to use this, and we've updated our internal Redux
subscription handling to leverage this API. This should also help
improve performance, by cutting down on the number of distinct renders
caused by a Redux store update.
function myThunk() {
return (dispatch, getState) => {
// should only result in one combined re-render, not two
batch(() => {
dispatch(increment());
dispatch(increment());
})
}
}
By design when the state, which is held by the store, changes the view should render.
You can avoid this by updating the state once.
If you are using promises you can use Promise.all to wait for all the promises to resolve and then dispatch a new action to the store with the calculated result. https://developer.mozilla.org/en/docs/Web/JavaScript/Reference/Global_Objects/Promise/all
Something like this:
Promise.all([p1, p2, p3, p4, p5]).then(changes => {
dispatch(changes)
}, err => {
// deal with error
});
Of course you'll need an action that will deal with many props, something like addManyProps this should update the state once, resulting in one render.
redux-batched-actions
Batching action creator and associated higher order reducer for redux that enables batching subscriber notifications for an array of actions.
Coming to this a bit late, but I think this is a much nicer solution, which enables you to add meta.batch to actions you would like to batch together into a single react update. As a bonus this approach works with asynchronous actions.
import raf from 'raf'
import { batchedSubscribe } from 'redux-batched-subscribe'
let notify = null
let rafId = null
const shouldBatch = action => action?.meta?.batch
export const batchedSubscribeEnhancer = batchedSubscribe(freshNotify => (notify = freshNotify))
export const batchedSubscribeMiddleware = () => next => action => {
const resolved = next(action)
if (notify && rafId === null && !shouldBatch(action)) {
notify()
} else if (!rafId) {
rafId = raf(() => {
rafId = null
notify()
})
}
return resolved
}
Then connect up to your store
mport { applyMiddleware, compose, createStore } from 'redux'
import { batchedSubscribeMiddleware, batchedSubscribeEnhancer } from './batching'
const store = createStore(
reducer,
intialState,
compose(
batchedSubscribeEnhancer,
applyMiddleware(batchedSubscribeMiddleware)
)
)
say I have a component connected to the redux store. Within this store, there's a list of objects. For instance something like this:
ReduxStore: {
dataList: [
{name:'bla'},
{name:'blub'},
]
}
Is it actually an anti pattern to create and add a filter function within the reducer to create something like this:
ReduxStore: {
dataList: {
data:[
{name:'bla'},
{name:'blub'}
],
isNameAvailable: (name) => {/* search for name */}
}
}
It works great, but I'm not sure whether this was the intended way to go.
It's an anti-pattern because your store only cares about data, not computation. Once you start adding functions to your store, you lose the ability to serialize the data inside it.
However, it's pretty trivial to pull these kinds of helper functions out and turn them into standalone utility selectors which you can use to achieve the same thing.
function isNameAvailable(store, name) {
/* search for `name` in `store` */
}
Now the function works independently of the store and you can keep them separate for testing.
From here, you can take a look at Reselect which allows you to turn your isNameAvailable function into a cached selector, meaning you'll only need to re-calculate as and when the appropriate data in the store is changed.
From Redux documentation:
It’s very important that the reducer stays pure. Things you should never do inside a reducer:
Mutate its arguments;
Perform side effects like API calls and routing transitions;
Call non-pure functions, e.g. Date.now() or Math.random().
so I think if you want to declare a filter function isNameAvailable, in my opinion, you have two options:
use the connect function to filter
// the solution of connect
import React, { Component } from 'react';
import { connect } from 'react-redux';
const App = () => (
<div>the example</div>
);
function isNameAvailable(state) {
// your code
}
function mapStateToProps(state) {
return isNameAvailable(state)
}
module.exports = connect(mapStateToProps)(App);
declare a filter function in action
In my experience,I think the first one is better. I hope this can help you.
Suppose a stateless, functional UserProfile component that displays user data for the given url. Suppose it is being wrapped with connect(mapStateToProps, mapDispatchToProps)(UserProfile). Finally, suppose a reducer that reduces into state.userProfile. Anytime the url changes, I need to re-initialize the state.userProfile, so a solution that comes to mind is to do so from within the mapDispatchToProps like so:
function mapDispatchToProps(dispatch, ownProps) {
dispatch(fetchUser(ownProps.userId))
return {
...
}
}
Provided that the thunked fetchUser ignores repeated calls by comparing with current state, is this an acceptable practice? Or are there problems associated with calling dispatch immediately from this map function?
This is unsupported and can break at any time.
mapDispatchToProps itself should not have side effects.
If you need to dispatch actions in response to prop changes, you can create a component class and use lifecycle methods for this:
class UserProfile extends Component {
componentDidMount() {
this.props.fetchUser(this.props.id)
}
componentDidUpdate(prevProps) {
if (prevProps.id !== this.props.id) {
this.props.fetchUser(this.props.id)
}
}
// ...
}