diff --git a/.gitignore b/.gitignore index 8996fca1..5a19a7f0 100644 --- a/.gitignore +++ b/.gitignore @@ -163,4 +163,7 @@ yarn-error.log* .playwright-mcp # YoYo AI version control directory -.yoyo/ \ No newline at end of file +.yoyo/ + +.specstory/ +.cursorindexingignore \ No newline at end of file diff --git a/docs/react-guide.md b/docs/react-guide.md new file mode 100644 index 00000000..957de419 --- /dev/null +++ b/docs/react-guide.md @@ -0,0 +1,813 @@ +# React guide + +A lot of AI-generated React code is heavily biased toward using useState and useEffect, often in verbose or naive ways. That’s because these hooks are: + +1. Explicit, making them easy for LLMs to pattern-match. +2. Safe defaults, requiring less architectural decision-making. +3. Well-documented, so the model’s training data is oversaturated with examples. + +But modern, well-architected React apps rarely rely on raw useState/useEffect beyond simple UI state or lifecycle quirks. Here’s a breakdown of what serious devs use instead—and why this is the real moat for human engineers. + +Alternatives to useState and useEffect + +1. State Machines / Statecharts (e.g. Zustand) +- Why: Replaces ad-hoc local state management with deterministic, declarative logic. +- Advantage: Less bugs, easier to test, clearer state transitions. +- AI gap: LLMs struggle to model state transitions declaratively. LLM Agent must overcome this. + +2. React Query / SWR for Data Fetching +- Why: Avoids useEffect-based data fetching and loading/error state juggling. +- Advantage: Caching, stale-while-revalidate, retries, pagination—all handled out of the box. +- AI gap: AI often uses useEffect + fetch and forgets cleanup, race conditions, or caching. + +3. Global State with Zustand, Jotai, Recoil, or Redux Toolkit +- Why: Local useState doesn’t scale for shared or persistent state. +- Advantage: Better performance, devtools support, persistence. +- AI gap: Defaults to prop-drilling or bloated context APIs. + +4. Custom Hooks and Composition +- Why: Reuse logic cleanly without bloating components or copy-pasting useEffect. +- Advantage: Separation of concerns, encapsulation. +- AI gap: Often fails to factor logic out, keeping everything in one component. + +## Why This Matters + +AI can mimic patterns, but it can’t: +- Architect a system for long-term maintainability. +- Predict runtime performance tradeoffs. +- Refactor spaghetti effects into custom hooks or state machines. +- Decide when state should be colocated vs globalized. +- Avoid footguns like stale closures, unnecessary re-renders, or race conditions. + +In short: AI is great at writing code, but bad at engineering software. + +Takeaway +- Use fewer useEffect and useState hooks. +- Learn to design systems, not just code components. +- Reach for declarative, composable patterns. +- Understand caching, reactivity, and scalability beyond local state. + +Here’s a side-by-side breakdown of common AI-generated React patterns (bad) vs what experienced engineers write (good) across 4 key areas: state management, data fetching, side effects, and logic reuse. + +## 1. State Management + +❌ Bad (AI-style useState) + +```typescript +function TodoApp() { + const [todos, setTodos] = useState([]); + const [input, setInput] = useState(''); + + const addTodo = () => { + setTodos([...todos, { text: input }]); + setInput(''); + }; + + return ( + <> + setInput(e.target.value)} /> + + + + ); +} +``` + +✅ Good (Zustand store, state extracted) + +```typescript +import { create } from 'zustand'; + +const useTodoStore = create((set) => ({ + todos: [], + input: '', + addTodo: () => set((state) => ({ + todos: [...state.todos, { text: state.input }], + input: '', + })), + setInput: (val) => set({ input: val }), +})); + +function TodoApp() { + const { todos, input, setInput, addTodo } = useTodoStore(); + + return ( + <> + setInput(e.target.value)} /> + + + + ); +} + +``` + +## 2. Data Fetching + +❌ Bad (AI-style useEffect) + +```typescript +function UserList() { + const [users, setUsers] = useState([]); + + useEffect(() => { + fetch('/api/users') + .then(res => res.json()) + .then(setUsers); + }, []); + + return ( + + ); +} +``` + +✅ Good (React Query) + +```typescript +import { useQuery } from '@tanstack/react-query'; + +function UserList() { + const { data: users = [], isLoading } = useQuery({ + queryKey: ['users'], + queryFn: () => fetch('/api/users').then(res => res.json()), + }); + + if (isLoading) return
Loading...
; + + return ( + + ); +} +``` + +## 3. Side Effects + +❌ Bad (naive effect with race condition) + +```typescript +function Search({ term }) { + const [results, setResults] = useState([]); + + useEffect(() => { + fetch(`/api/search?q=${term}`) + .then(res => res.json()) + .then(setResults); + }, [term]); + + return
{results.length} results
; +} +``` + +✅ Good (abort previous fetch with AbortController) + +```typescript +function Search({ term }) { + const [results, setResults] = useState([]); + + useEffect(() => { + const controller = new AbortController(); + + fetch(`/api/search?q=${term}`, { signal: controller.signal }) + .then(res => res.json()) + .then(setResults) + .catch((e) => { + if (e.name !== 'AbortError') console.error(e); + }); + + return () => controller.abort(); + }, [term]); + + return
{results.length} results
; +} +``` + +Better yet? Don’t use useEffect at all. Use React Query with term as a key. + +## 4. Logic Reuse + +❌ Bad (repeated logic inline) + +```typescript +function ComponentA() { + const [count, setCount] = useState(0); + useEffect(() => { + const id = setInterval(() => setCount(c => c + 1), 1000); + return () => clearInterval(id); + }, []); + return
{count}
; +} + +function ComponentB() { + const [count, setCount] = useState(0); + useEffect(() => { + const id = setInterval(() => setCount(c => c + 1), 1000); + return () => clearInterval(id); + }, []); + return
{count}
; +} +``` + +✅ Good (custom hook) + +```typescript +function useCounter(intervalMs = 1000) { + const [count, setCount] = useState(0); + useEffect(() => { + const id = setInterval(() => setCount(c => c + 1), intervalMs); + return () => clearInterval(id); + }, [intervalMs]); + return count; +} + +function ComponentA() { + const count = useCounter(); + return
{count}
; +} + +function ComponentB() { + const count = useCounter(); + return
{count}
; +} + +``` + +## 5. State Machines with Zustand + +❌ Bad (AI-style complex useState with boolean flags) + +```typescript +function OrderProcess() { + const [isLoading, setIsLoading] = useState(false); + const [isError, setIsError] = useState(false); + const [isSuccess, setIsSuccess] = useState(false); + const [cart, setCart] = useState([]); + const [paymentDetails, setPaymentDetails] = useState(null); + const [orderPlaced, setOrderPlaced] = useState(false); + + const addToCart = (item) => { + setCart([...cart, item]); + }; + + const submitPayment = async (details) => { + setIsLoading(true); + setIsError(false); + + try { + await processPayment(details); + setPaymentDetails(details); + setIsSuccess(true); + setIsLoading(false); + } catch (error) { + setIsError(true); + setIsLoading(false); + } + }; + + const placeOrder = async () => { + setIsLoading(true); + setIsError(false); + + try { + await submitOrder(cart, paymentDetails); + setOrderPlaced(true); + setIsSuccess(true); + setIsLoading(false); + } catch (error) { + setIsError(true); + setIsLoading(false); + } + }; + + return ( +
+ {/* Complex conditional rendering based on multiple state variables */} + {isLoading && } + {isError && } + {!isLoading && !isError && !orderPlaced && ( + <> + + + {paymentDetails && } + + )} + {orderPlaced && } +
+ ); +} +``` + +✅ Good (Zustand state machine) + +```typescript +import { create } from 'zustand'; + +// Create a Zustand store with state machine pattern +const useOrderStore = create((set, get) => ({ + // State machine's current state + state: 'browsing', + + // Context/data + cart: [], + paymentDetails: null, + error: null, + + // Actions that transition between states + addToCart: (item) => set((state) => ({ + cart: [...state.cart, item] + })), + + goToCheckout: () => set({ state: 'paymentEntry' }), + + goBack: () => { + const { state: currentState } = get(); + if (currentState === 'paymentEntry') { + set({ state: 'browsing' }); + } else if (currentState === 'confirmOrder') { + set({ state: 'paymentEntry' }); + } + }, + + submitPayment: async (details) => { + set({ state: 'processingPayment', error: null }); + + try { + const result = await processPayment(details); + set({ + state: 'confirmOrder', + paymentDetails: result + }); + } catch (error) { + set({ + state: 'paymentEntry', + error: error.message + }); + } + }, + + placeOrder: async () => { + const { cart, paymentDetails } = get(); + set({ state: 'processingOrder', error: null }); + + try { + await submitOrder(cart, paymentDetails); + set({ state: 'orderComplete' }); + } catch (error) { + set({ + state: 'confirmOrder', + error: error.message + }); + } + } +})); + +function OrderProcess() { + const { + state, + cart, + paymentDetails, + error, + addToCart, + goToCheckout, + goBack, + submitPayment, + placeOrder + } = useOrderStore(); + + // Render UI based on current state + return ( +
+ {error && } + + {state === 'browsing' && ( + + )} + + {state === 'paymentEntry' && ( + + )} + + {state === 'processingPayment' && ( + + )} + + {state === 'confirmOrder' && ( + + )} + + {state === 'processingOrder' && ( + + )} + + {state === 'orderComplete' && ( + + )} +
+ ); +} +``` + +### Key Benefits of the Zustand State Machine Approach: + +1. **Explicit states**: The system can only be in one well-defined state at a time +2. **Centralized logic**: All state transitions are defined in one place +3. **Predictable transitions**: State changes follow clear patterns +4. **Simplified UI logic**: Components just render based on the current state +5. **Async handling**: Asynchronous operations are contained within the state transitions +6. **Shared state**: Any component can access the store without prop drilling +7. **Improved debugging**: The current state is always clear and predictable +8. **No impossible states**: Unlike boolean flags that could create invalid combinations +9. **Simpler API**: Compared to XState, Zustand offers a more approachable API for many developers +10. **Smaller bundle size**: Zustand is significantly smaller than XState + +This approach gives most of the benefits of formal state machines while maintaining the simplicity and familiarity of React state management. It's a pragmatic middle ground that's often easier to adopt in existing projects. + +## 6. Concurrent UI Techniques + +❌ Bad (Blocking UI updates without concurrency): +```jsx +// Bad example: Synchronously fetching data in useEffect can cause UI jank. +function SearchResults({ query }) { + const [results, setResults] = React.useState([]); + + React.useEffect(() => { + fetch(`/api/search?q=${query}`) + .then(res => res.json()) + .then(data => setResults(data)); + }, [query]); + + return ( +
+ {results.map(item =>

{item.title}

)} +
+ ); +} +``` + +✅ Good (Using useTransition for non-blocking updates): +```jsx +// Good example: Leveraging useTransition to mark state updates as low-priority. +function SearchResults({ query }) { + const [results, setResults] = React.useState([]); + const [isPending, startTransition] = React.useTransition(); + + React.useEffect(() => { + startTransition(() => { + fetch(`/api/search?q=${query}`) + .then(res => res.json()) + .then(data => setResults(data)); + }); + }, [query, startTransition]); + + return ( +
+ {isPending &&
Loading...
} + {results.map(item =>

{item.title}

)} +
+ ); +} +``` + +## 7. Routing and Navigation (Not applicable with Next.js apps) + +❌ Bad (Manual routing without a dedicated library): +```jsx +// Bad example: Handling routing manually using state and window history. +function App() { + const [route, setRoute] = React.useState(window.location.pathname); + + React.useEffect(() => { + const onPopState = () => setRoute(window.location.pathname); + window.addEventListener('popstate', onPopState); + return () => window.removeEventListener('popstate', onPopState); + }, []); + + return ( +
+ {route === '/' && } + {route === '/about' && } + +
+ ); +} +``` + +✅ Good (Using React Router v6 with lazy-loaded routes): +```jsx +// Good example: Utilizing React Router to manage routes and lazy load components. +import { BrowserRouter, Routes, Route, Link } from 'react-router-dom'; +import React, { lazy, Suspense } from 'react'; + +const Home = lazy(() => import('./Home')); +const About = lazy(() => import('./About')); + +function App() { + return ( + + + Loading...}> + + } /> + } /> + + + + ); +} +``` + +## 8. Error Handling + +❌ Bad (No error boundary; crashes on errors): +```jsx +// Bad example: A component that throws errors without an error boundary. +function MyComponent({ data }) { + if (!data) { + throw new Error("Data not found"); + } + return
{data}
; +} + +function App() { + return ; +} +``` + +✅ Good (Wrapping components with an Error Boundary): +```jsx +// Good example: Using an ErrorBoundary component to catch and handle errors gracefully. +import React from 'react'; + +class ErrorBoundary extends React.Component { + constructor(props) { + super(props); + this.state = { hasError: false }; + } + + static getDerivedStateFromError(error) { + return { hasError: true }; + } + + componentDidCatch(error, errorInfo) { + console.error("Error caught in boundary:", error, errorInfo); + } + + render() { + if (this.state.hasError) { + return

Something went wrong.

; + } + return this.props.children; + } +} + +function MyComponent({ data }) { + if (!data) { + throw new Error("Data not found"); + } + return
{data}
; +} + +function App() { + return ( + + + + ); +} +``` + +## 9. Form Management + +❌ Bad (Manual state handling with no validation): +```jsx +// Bad example: Uncontrolled form without validation; error handling is omitted. +function ContactForm() { + const [formData, setFormData] = React.useState({ name: '', email: '' }); + + const handleSubmit = (e) => { + e.preventDefault(); + console.log(formData); + }; + + return ( +
+ setFormData({ ...formData, name: e.target.value })} + placeholder="Name" + /> + setFormData({ ...formData, email: e.target.value })} + placeholder="Email" + /> + +
+ ); +} +``` + +✅ Good (Using React Hook Form with Zod for schema validation): + +```jsx +import React from 'react'; +import { useForm } from 'react-hook-form'; +import { zodResolver } from '@hookform/resolvers/zod'; +import { z } from 'zod'; + +// 1. Define your schema with Zod +const contactSchema = z.object({ + name: z.string().nonempty('Name is required'), + email: z + .string() + .nonempty('Email is required') + .email('Invalid email'), +}); + +// 2. Infer the TypeScript type of your form data +type ContactFormData = z.infer; + +export function ContactForm() { + // 3. Hook up RHF with the zodResolver + const { + register, + handleSubmit, + formState: { errors }, + } = useForm({ + resolver: zodResolver(contactSchema), + }); + + const onSubmit = (data: ContactFormData) => { + console.log(data); + }; + + return ( +
+
+ + {errors.name &&

{errors.name.message}

} +
+ +
+ + {errors.email &&

{errors.email.message}

} +
+ + +
+ ); +} +``` + +## 10. Code Splitting and Lazy Loading + +❌ Bad (Eagerly importing heavy components, increasing the initial bundle size): +```jsx +// Bad example: Directly importing a large component causing a heavier bundle. +import HeavyComponent from './HeavyComponent'; + +function App() { + return ( +
+ +
+ ); +} +``` + +✅ Good (Lazy loading heavy components with React.lazy and Suspense): +```jsx +// Good example: Dynamically importing components to reduce initial load time. +import React, { lazy, Suspense } from 'react'; + +const HeavyComponent = lazy(() => import('./HeavyComponent')); + +function App() { + return ( + Loading Component...}> + + + ); +} +``` + +## 11. Testing Strategies + +❌ Bad (Testing implementation details, resulting in brittle tests): +```jsx +// Bad example: Relying on internal DOM structure which may change. +import { render } from '@testing-library/react'; +import MyComponent from './MyComponent'; + +test('MyComponent renders data', () => { + const { container } = render(); + expect(container.querySelector('.data')).toBeDefined(); +}); +``` + +✅ Good (Focusing on user-centric behaviors using React Testing Library): +```jsx +// Good example: Testing component behavior by simulating user interactions. +import { render, screen } from '@testing-library/react'; +import userEvent from '@testing-library/user-event'; +import MyComponent from './MyComponent'; + +test('MyComponent displays loaded data after user action', async () => { + render(); + + // Check for an initial loading state + expect(screen.getByText(/Loading/i)).toBeInTheDocument(); + + // Simulate user action that triggers data loading + const button = screen.getByRole('button', { name: /load data/i }); + userEvent.click(button); + + // Wait and assert that data is displayed + expect(await screen.findByText(/Data loaded/i)).toBeInTheDocument(); +}); +``` + +## 12. Accessibility & Theming + +❌ Bad (Non-semantic elements and missing ARIA labels): +```jsx +// Bad example: Using a div as a clickable element without semantics or accessibility. +function ThemedButton() { + return
console.log('Clicked!')}>Click me
; +} +``` + +✅ Good (Semantic button with ARIA attributes and theming via CSS variables): +```jsx +// Good example: Accessible button component that utilizes theming. +function ThemedButton() { + return ( + + ); +} + +/* CSS (in a separate file or styled-components): +:root { + --primary-color: #007bff; +} +*/ +``` + +# Summary Table + +| Concern | Bad AI Pattern | Solid Engineering Practice | +|---------|---------------|----------------------------| +| **State Management** | `useState` everywhere, prop‑drilling | Central store (Zustand / Redux Toolkit) or colocated custom hooks | +| **Data Fetching** | `useEffect` + `fetch`, manual loading/error juggling | React Query / SWR with caching, retries, stale‑while‑revalidate | +| **Side Effects** | Effects without cleanup → race conditions, leaks | AbortController or query library handles cancellation & retries | +| **Logic Reuse** | Copy‑pasted hooks and state in every component | Extract cross‑cutting logic into custom hooks | +| **Workflow / State Machine** | Boolean‑flag soup, impossible states | Deterministic statechart in a single store (Zustand / XState) | +| **Concurrent UI Updates** | Blocking synchronous work inside effects | `useTransition` / `startTransition` for low‑priority updates | +| **Routing & Navigation** | DIY history manipulation, inline route state | React Router v6 (or Next.js built‑ins) with lazy‑loaded routes | +| **Error Handling** | No error boundary → full‑app crashes | Top‑level `ErrorBoundary` plus logging in `componentDidCatch` | +| **Form Management** | Manual state, zero validation or feedback | React Hook Form + Yup/Zod schema for declarative validation | +| **Code Splitting & Lazy Loading** | Eager imports ballooning bundle size | `React.lazy` + `Suspense` for on‑demand loading | +| **Testing Strategy** | Brittle tests keyed to DOM structure | User‑centric tests with React Testing Library & jest‑dom | +| **Accessibility & Theming** | Non‑semantic elements, missing ARIA, inline colors | Semantic HTML, ARIA labels, CSS variables / theme provider | \ No newline at end of file diff --git a/docs/you-might-not-need-an-effect.md b/docs/you-might-not-need-an-effect.md new file mode 100644 index 00000000..fa91579f --- /dev/null +++ b/docs/you-might-not-need-an-effect.md @@ -0,0 +1,1745 @@ +--- +title: 'You Might Not Need an Effect' +source: https://react.dev/learn/you-might-not-need-an-effect +--- + + + +Effects are an escape hatch from the React paradigm. They let you "step outside" of React and synchronize your components with some external system like a non-React widget, network, or the browser DOM. If there is no external system involved (for example, if you want to update a component's state when some props or state change), you shouldn't need an Effect. Removing unnecessary Effects will make your code easier to follow, faster to run, and less error-prone. + + + + + +* Why and how to remove unnecessary Effects from your components +* How to cache expensive computations without Effects +* How to reset and adjust component state without Effects +* How to share logic between event handlers +* Which logic should be moved to event handlers +* How to notify parent components about changes + + + +## How to remove unnecessary Effects {/*how-to-remove-unnecessary-effects*/} + +There are two common cases in which you don't need Effects: + +* **You don't need Effects to transform data for rendering.** For example, let's say you want to filter a list before displaying it. You might feel tempted to write an Effect that updates a state variable when the list changes. However, this is inefficient. When you update the state, React will first call your component functions to calculate what should be on the screen. Then React will ["commit"](/learn/render-and-commit) these changes to the DOM, updating the screen. Then React will run your Effects. If your Effect *also* immediately updates the state, this restarts the whole process from scratch! To avoid the unnecessary render passes, transform all the data at the top level of your components. That code will automatically re-run whenever your props or state change. +* **You don't need Effects to handle user events.** For example, let's say you want to send an `/api/buy` POST request and show a notification when the user buys a product. In the Buy button click event handler, you know exactly what happened. By the time an Effect runs, you don't know *what* the user did (for example, which button was clicked). This is why you'll usually handle user events in the corresponding event handlers. + +You *do* need Effects to [synchronize](/learn/synchronizing-with-effects#what-are-effects-and-how-are-they-different-from-events) with external systems. For example, you can write an Effect that keeps a jQuery widget synchronized with the React state. You can also fetch data with Effects: for example, you can synchronize the search results with the current search query. Keep in mind that modern [frameworks](/learn/start-a-new-react-project#full-stack-frameworks) provide more efficient built-in data fetching mechanisms than writing Effects directly in your components. + +To help you gain the right intuition, let's look at some common concrete examples! + +### Updating state based on props or state {/*updating-state-based-on-props-or-state*/} + +Suppose you have a component with two state variables: `firstName` and `lastName`. You want to calculate a `fullName` from them by concatenating them. Moreover, you'd like `fullName` to update whenever `firstName` or `lastName` change. Your first instinct might be to add a `fullName` state variable and update it in an Effect: + +```js {5-9} +function Form() { + const [firstName, setFirstName] = useState('Taylor'); + const [lastName, setLastName] = useState('Swift'); + + // 🔴 Avoid: redundant state and unnecessary Effect + const [fullName, setFullName] = useState(''); + useEffect(() => { + setFullName(firstName + ' ' + lastName); + }, [firstName, lastName]); + // ... +} +``` + +This is more complicated than necessary. It is inefficient too: it does an entire render pass with a stale value for `fullName`, then immediately re-renders with the updated value. Remove the state variable and the Effect: + +```js {4-5} +function Form() { + const [firstName, setFirstName] = useState('Taylor'); + const [lastName, setLastName] = useState('Swift'); + // ✅ Good: calculated during rendering + const fullName = firstName + ' ' + lastName; + // ... +} +``` + +**When something can be calculated from the existing props or state, [don't put it in state.](/learn/choosing-the-state-structure#avoid-redundant-state) Instead, calculate it during rendering.** This makes your code faster (you avoid the extra "cascading" updates), simpler (you remove some code), and less error-prone (you avoid bugs caused by different state variables getting out of sync with each other). If this approach feels new to you, [Thinking in React](/learn/thinking-in-react#step-3-find-the-minimal-but-complete-representation-of-ui-state) explains what should go into state. + +### Caching expensive calculations {/*caching-expensive-calculations*/} + +This component computes `visibleTodos` by taking the `todos` it receives by props and filtering them according to the `filter` prop. You might feel tempted to store the result in state and update it from an Effect: + +```js {4-8} +function TodoList({ todos, filter }) { + const [newTodo, setNewTodo] = useState(''); + + // 🔴 Avoid: redundant state and unnecessary Effect + const [visibleTodos, setVisibleTodos] = useState([]); + useEffect(() => { + setVisibleTodos(getFilteredTodos(todos, filter)); + }, [todos, filter]); + + // ... +} +``` + +Like in the earlier example, this is both unnecessary and inefficient. First, remove the state and the Effect: + +```js {3-4} +function TodoList({ todos, filter }) { + const [newTodo, setNewTodo] = useState(''); + // ✅ This is fine if getFilteredTodos() is not slow. + const visibleTodos = getFilteredTodos(todos, filter); + // ... +} +``` + +Usually, this code is fine! But maybe `getFilteredTodos()` is slow or you have a lot of `todos`. In that case you don't want to recalculate `getFilteredTodos()` if some unrelated state variable like `newTodo` has changed. + +You can cache (or ["memoize"](https://en.wikipedia.org/wiki/Memoization)) an expensive calculation by wrapping it in a [`useMemo`](/reference/react/useMemo) Hook: + + + +[React Compiler](/learn/react-compiler) can automatically memoize expensive calculations for you, eliminating the need for manual `useMemo` in many cases. + + + +```js {5-8} +import { useMemo, useState } from 'react'; + +function TodoList({ todos, filter }) { + const [newTodo, setNewTodo] = useState(''); + const visibleTodos = useMemo(() => { + // ✅ Does not re-run unless todos or filter change + return getFilteredTodos(todos, filter); + }, [todos, filter]); + // ... +} +``` + +Or, written as a single line: + +```js {5-6} +import { useMemo, useState } from 'react'; + +function TodoList({ todos, filter }) { + const [newTodo, setNewTodo] = useState(''); + // ✅ Does not re-run getFilteredTodos() unless todos or filter change + const visibleTodos = useMemo(() => getFilteredTodos(todos, filter), [todos, filter]); + // ... +} +``` + +**This tells React that you don't want the inner function to re-run unless either `todos` or `filter` have changed.** React will remember the return value of `getFilteredTodos()` during the initial render. During the next renders, it will check if `todos` or `filter` are different. If they're the same as last time, `useMemo` will return the last result it has stored. But if they are different, React will call the inner function again (and store its result). + +The function you wrap in [`useMemo`](/reference/react/useMemo) runs during rendering, so this only works for [pure calculations.](/learn/keeping-components-pure) + + + +#### How to tell if a calculation is expensive? {/*how-to-tell-if-a-calculation-is-expensive*/} + +In general, unless you're creating or looping over thousands of objects, it's probably not expensive. If you want to get more confidence, you can add a console log to measure the time spent in a piece of code: + +```js {1,3} +console.time('filter array'); +const visibleTodos = getFilteredTodos(todos, filter); +console.timeEnd('filter array'); +``` + +Perform the interaction you're measuring (for example, typing into the input). You will then see logs like `filter array: 0.15ms` in your console. If the overall logged time adds up to a significant amount (say, `1ms` or more), it might make sense to memoize that calculation. As an experiment, you can then wrap the calculation in `useMemo` to verify whether the total logged time has decreased for that interaction or not: + +```js +console.time('filter array'); +const visibleTodos = useMemo(() => { + return getFilteredTodos(todos, filter); // Skipped if todos and filter haven't changed +}, [todos, filter]); +console.timeEnd('filter array'); +``` + +`useMemo` won't make the *first* render faster. It only helps you skip unnecessary work on updates. + +Keep in mind that your machine is probably faster than your users' so it's a good idea to test the performance with an artificial slowdown. For example, Chrome offers a [CPU Throttling](https://developer.chrome.com/blog/new-in-devtools-61/#throttling) option for this. + +Also note that measuring performance in development will not give you the most accurate results. (For example, when [Strict Mode](/reference/react/StrictMode) is on, you will see each component render twice rather than once.) To get the most accurate timings, build your app for production and test it on a device like your users have. + + + +### Resetting all state when a prop changes {/*resetting-all-state-when-a-prop-changes*/} + +This `ProfilePage` component receives a `userId` prop. The page contains a comment input, and you use a `comment` state variable to hold its value. One day, you notice a problem: when you navigate from one profile to another, the `comment` state does not get reset. As a result, it's easy to accidentally post a comment on a wrong user's profile. To fix the issue, you want to clear out the `comment` state variable whenever the `userId` changes: + +```js {4-7} +export default function ProfilePage({ userId }) { + const [comment, setComment] = useState(''); + + // 🔴 Avoid: Resetting state on prop change in an Effect + useEffect(() => { + setComment(''); + }, [userId]); + // ... +} +``` + +This is inefficient because `ProfilePage` and its children will first render with the stale value, and then render again. It is also complicated because you'd need to do this in *every* component that has some state inside `ProfilePage`. For example, if the comment UI is nested, you'd want to clear out nested comment state too. + +Instead, you can tell React that each user's profile is conceptually a _different_ profile by giving it an explicit key. Split your component in two and pass a `key` attribute from the outer component to the inner one: + +```js {5,11-12} +export default function ProfilePage({ userId }) { + return ( + + ); +} + +function Profile({ userId }) { + // ✅ This and any other state below will reset on key change automatically + const [comment, setComment] = useState(''); + // ... +} +``` + +Normally, React preserves the state when the same component is rendered in the same spot. **By passing `userId` as a `key` to the `Profile` component, you're asking React to treat two `Profile` components with different `userId` as two different components that should not share any state.** Whenever the key (which you've set to `userId`) changes, React will recreate the DOM and [reset the state](/learn/preserving-and-resetting-state#option-2-resetting-state-with-a-key) of the `Profile` component and all of its children. Now the `comment` field will clear out automatically when navigating between profiles. + +Note that in this example, only the outer `ProfilePage` component is exported and visible to other files in the project. Components rendering `ProfilePage` don't need to pass the key to it: they pass `userId` as a regular prop. The fact `ProfilePage` passes it as a `key` to the inner `Profile` component is an implementation detail. + +### Adjusting some state when a prop changes {/*adjusting-some-state-when-a-prop-changes*/} + +Sometimes, you might want to reset or adjust a part of the state on a prop change, but not all of it. + +This `List` component receives a list of `items` as a prop, and maintains the selected item in the `selection` state variable. You want to reset the `selection` to `null` whenever the `items` prop receives a different array: + +```js {5-8} +function List({ items }) { + const [isReverse, setIsReverse] = useState(false); + const [selection, setSelection] = useState(null); + + // 🔴 Avoid: Adjusting state on prop change in an Effect + useEffect(() => { + setSelection(null); + }, [items]); + // ... +} +``` + +This, too, is not ideal. Every time the `items` change, the `List` and its child components will render with a stale `selection` value at first. Then React will update the DOM and run the Effects. Finally, the `setSelection(null)` call will cause another re-render of the `List` and its child components, restarting this whole process again. + +Start by deleting the Effect. Instead, adjust the state directly during rendering: + +```js {5-11} +function List({ items }) { + const [isReverse, setIsReverse] = useState(false); + const [selection, setSelection] = useState(null); + + // Better: Adjust the state while rendering + const [prevItems, setPrevItems] = useState(items); + if (items !== prevItems) { + setPrevItems(items); + setSelection(null); + } + // ... +} +``` + +[Storing information from previous renders](/reference/react/useState#storing-information-from-previous-renders) like this can be hard to understand, but it’s better than updating the same state in an Effect. In the above example, `setSelection` is called directly during a render. React will re-render the `List` *immediately* after it exits with a `return` statement. React has not rendered the `List` children or updated the DOM yet, so this lets the `List` children skip rendering the stale `selection` value. + +When you update a component during rendering, React throws away the returned JSX and immediately retries rendering. To avoid very slow cascading retries, React only lets you update the *same* component's state during a render. If you update another component's state during a render, you'll see an error. A condition like `items !== prevItems` is necessary to avoid loops. You may adjust state like this, but any other side effects (like changing the DOM or setting timeouts) should stay in event handlers or Effects to [keep components pure.](/learn/keeping-components-pure) + +**Although this pattern is more efficient than an Effect, most components shouldn't need it either.** No matter how you do it, adjusting state based on props or other state makes your data flow more difficult to understand and debug. Always check whether you can [reset all state with a key](#resetting-all-state-when-a-prop-changes) or [calculate everything during rendering](#updating-state-based-on-props-or-state) instead. For example, instead of storing (and resetting) the selected *item*, you can store the selected *item ID:* + +```js {3-5} +function List({ items }) { + const [isReverse, setIsReverse] = useState(false); + const [selectedId, setSelectedId] = useState(null); + // ✅ Best: Calculate everything during rendering + const selection = items.find(item => item.id === selectedId) ?? null; + // ... +} +``` + +Now there is no need to "adjust" the state at all. If the item with the selected ID is in the list, it remains selected. If it's not, the `selection` calculated during rendering will be `null` because no matching item was found. This behavior is different, but arguably better because most changes to `items` preserve the selection. + +### Sharing logic between event handlers {/*sharing-logic-between-event-handlers*/} + +Let's say you have a product page with two buttons (Buy and Checkout) that both let you buy that product. You want to show a notification whenever the user puts the product in the cart. Calling `showNotification()` in both buttons' click handlers feels repetitive so you might be tempted to place this logic in an Effect: + +```js {2-7} +function ProductPage({ product, addToCart }) { + // 🔴 Avoid: Event-specific logic inside an Effect + useEffect(() => { + if (product.isInCart) { + showNotification(`Added ${product.name} to the shopping cart!`); + } + }, [product]); + + function handleBuyClick() { + addToCart(product); + } + + function handleCheckoutClick() { + addToCart(product); + navigateTo('/checkout'); + } + // ... +} +``` + +This Effect is unnecessary. It will also most likely cause bugs. For example, let's say that your app "remembers" the shopping cart between the page reloads. If you add a product to the cart once and refresh the page, the notification will appear again. It will keep appearing every time you refresh that product's page. This is because `product.isInCart` will already be `true` on the page load, so the Effect above will call `showNotification()`. + +**When you're not sure whether some code should be in an Effect or in an event handler, ask yourself *why* this code needs to run. Use Effects only for code that should run *because* the component was displayed to the user.** In this example, the notification should appear because the user *pressed the button*, not because the page was displayed! Delete the Effect and put the shared logic into a function called from both event handlers: + +```js {2-6,9,13} +function ProductPage({ product, addToCart }) { + // ✅ Good: Event-specific logic is called from event handlers + function buyProduct() { + addToCart(product); + showNotification(`Added ${product.name} to the shopping cart!`); + } + + function handleBuyClick() { + buyProduct(); + } + + function handleCheckoutClick() { + buyProduct(); + navigateTo('/checkout'); + } + // ... +} +``` + +This both removes the unnecessary Effect and fixes the bug. + +### Sending a POST request {/*sending-a-post-request*/} + +This `Form` component sends two kinds of POST requests. It sends an analytics event when it mounts. When you fill in the form and click the Submit button, it will send a POST request to the `/api/register` endpoint: + +```js {5-8,10-16} +function Form() { + const [firstName, setFirstName] = useState(''); + const [lastName, setLastName] = useState(''); + + // ✅ Good: This logic should run because the component was displayed + useEffect(() => { + post('/analytics/event', { eventName: 'visit_form' }); + }, []); + + // 🔴 Avoid: Event-specific logic inside an Effect + const [jsonToSubmit, setJsonToSubmit] = useState(null); + useEffect(() => { + if (jsonToSubmit !== null) { + post('/api/register', jsonToSubmit); + } + }, [jsonToSubmit]); + + function handleSubmit(e) { + e.preventDefault(); + setJsonToSubmit({ firstName, lastName }); + } + // ... +} +``` + +Let's apply the same criteria as in the example before. + +The analytics POST request should remain in an Effect. This is because the _reason_ to send the analytics event is that the form was displayed. (It would fire twice in development, but [see here](/learn/synchronizing-with-effects#sending-analytics) for how to deal with that.) + +However, the `/api/register` POST request is not caused by the form being _displayed_. You only want to send the request at one specific moment in time: when the user presses the button. It should only ever happen _on that particular interaction_. Delete the second Effect and move that POST request into the event handler: + +```js {12-13} +function Form() { + const [firstName, setFirstName] = useState(''); + const [lastName, setLastName] = useState(''); + + // ✅ Good: This logic runs because the component was displayed + useEffect(() => { + post('/analytics/event', { eventName: 'visit_form' }); + }, []); + + function handleSubmit(e) { + e.preventDefault(); + // ✅ Good: Event-specific logic is in the event handler + post('/api/register', { firstName, lastName }); + } + // ... +} +``` + +When you choose whether to put some logic into an event handler or an Effect, the main question you need to answer is _what kind of logic_ it is from the user's perspective. If this logic is caused by a particular interaction, keep it in the event handler. If it's caused by the user _seeing_ the component on the screen, keep it in the Effect. + +### Chains of computations {/*chains-of-computations*/} + +Sometimes you might feel tempted to chain Effects that each adjust a piece of state based on other state: + +```js {7-29} +function Game() { + const [card, setCard] = useState(null); + const [goldCardCount, setGoldCardCount] = useState(0); + const [round, setRound] = useState(1); + const [isGameOver, setIsGameOver] = useState(false); + + // 🔴 Avoid: Chains of Effects that adjust the state solely to trigger each other + useEffect(() => { + if (card !== null && card.gold) { + setGoldCardCount(c => c + 1); + } + }, [card]); + + useEffect(() => { + if (goldCardCount > 3) { + setRound(r => r + 1) + setGoldCardCount(0); + } + }, [goldCardCount]); + + useEffect(() => { + if (round > 5) { + setIsGameOver(true); + } + }, [round]); + + useEffect(() => { + alert('Good game!'); + }, [isGameOver]); + + function handlePlaceCard(nextCard) { + if (isGameOver) { + throw Error('Game already ended.'); + } else { + setCard(nextCard); + } + } + + // ... +``` + +There are two problems with this code. + +The first problem is that it is very inefficient: the component (and its children) have to re-render between each `set` call in the chain. In the example above, in the worst case (`setCard` → render → `setGoldCardCount` → render → `setRound` → render → `setIsGameOver` → render) there are three unnecessary re-renders of the tree below. + +The second problem is that even if it weren't slow, as your code evolves, you will run into cases where the "chain" you wrote doesn't fit the new requirements. Imagine you are adding a way to step through the history of the game moves. You'd do it by updating each state variable to a value from the past. However, setting the `card` state to a value from the past would trigger the Effect chain again and change the data you're showing. Such code is often rigid and fragile. + +In this case, it's better to calculate what you can during rendering, and adjust the state in the event handler: + +```js {6-7,14-26} +function Game() { + const [card, setCard] = useState(null); + const [goldCardCount, setGoldCardCount] = useState(0); + const [round, setRound] = useState(1); + + // ✅ Calculate what you can during rendering + const isGameOver = round > 5; + + function handlePlaceCard(nextCard) { + if (isGameOver) { + throw Error('Game already ended.'); + } + + // ✅ Calculate all the next state in the event handler + setCard(nextCard); + if (nextCard.gold) { + if (goldCardCount <= 3) { + setGoldCardCount(goldCardCount + 1); + } else { + setGoldCardCount(0); + setRound(round + 1); + if (round === 5) { + alert('Good game!'); + } + } + } + } + + // ... +``` + +This is a lot more efficient. Also, if you implement a way to view game history, now you will be able to set each state variable to a move from the past without triggering the Effect chain that adjusts every other value. If you need to reuse logic between several event handlers, you can [extract a function](#sharing-logic-between-event-handlers) and call it from those handlers. + +Remember that inside event handlers, [state behaves like a snapshot.](/learn/state-as-a-snapshot) For example, even after you call `setRound(round + 1)`, the `round` variable will reflect the value at the time the user clicked the button. If you need to use the next value for calculations, define it manually like `const nextRound = round + 1`. + +In some cases, you *can't* calculate the next state directly in the event handler. For example, imagine a form with multiple dropdowns where the options of the next dropdown depend on the selected value of the previous dropdown. Then, a chain of Effects is appropriate because you are synchronizing with network. + +### Initializing the application {/*initializing-the-application*/} + +Some logic should only run once when the app loads. + +You might be tempted to place it in an Effect in the top-level component: + +```js {2-6} +function App() { + // 🔴 Avoid: Effects with logic that should only ever run once + useEffect(() => { + loadDataFromLocalStorage(); + checkAuthToken(); + }, []); + // ... +} +``` + +However, you'll quickly discover that it [runs twice in development.](/learn/synchronizing-with-effects#how-to-handle-the-effect-firing-twice-in-development) This can cause issues--for example, maybe it invalidates the authentication token because the function wasn't designed to be called twice. In general, your components should be resilient to being remounted. This includes your top-level `App` component. + +Although it may not ever get remounted in practice in production, following the same constraints in all components makes it easier to move and reuse code. If some logic must run *once per app load* rather than *once per component mount*, add a top-level variable to track whether it has already executed: + +```js {1,5-6,10} +let didInit = false; + +function App() { + useEffect(() => { + if (!didInit) { + didInit = true; + // ✅ Only runs once per app load + loadDataFromLocalStorage(); + checkAuthToken(); + } + }, []); + // ... +} +``` + +You can also run it during module initialization and before the app renders: + +```js {1,5} +if (typeof window !== 'undefined') { // Check if we're running in the browser. + // ✅ Only runs once per app load + checkAuthToken(); + loadDataFromLocalStorage(); +} + +function App() { + // ... +} +``` + +Code at the top level runs once when your component is imported--even if it doesn't end up being rendered. To avoid slowdown or surprising behavior when importing arbitrary components, don't overuse this pattern. Keep app-wide initialization logic to root component modules like `App.js` or in your application's entry point. + +### Notifying parent components about state changes {/*notifying-parent-components-about-state-changes*/} + +Let's say you're writing a `Toggle` component with an internal `isOn` state which can be either `true` or `false`. There are a few different ways to toggle it (by clicking or dragging). You want to notify the parent component whenever the `Toggle` internal state changes, so you expose an `onChange` event and call it from an Effect: + +```js {4-7} +function Toggle({ onChange }) { + const [isOn, setIsOn] = useState(false); + + // 🔴 Avoid: The onChange handler runs too late + useEffect(() => { + onChange(isOn); + }, [isOn, onChange]) + + function handleClick() { + setIsOn(!isOn); + } + + function handleDragEnd(e) { + if (isCloserToRightEdge(e)) { + setIsOn(true); + } else { + setIsOn(false); + } + } + + // ... +} +``` + +Like earlier, this is not ideal. The `Toggle` updates its state first, and React updates the screen. Then React runs the Effect, which calls the `onChange` function passed from a parent component. Now the parent component will update its own state, starting another render pass. It would be better to do everything in a single pass. + +Delete the Effect and instead update the state of *both* components within the same event handler: + +```js {5-7,11,16,18} +function Toggle({ onChange }) { + const [isOn, setIsOn] = useState(false); + + function updateToggle(nextIsOn) { + // ✅ Good: Perform all updates during the event that caused them + setIsOn(nextIsOn); + onChange(nextIsOn); + } + + function handleClick() { + updateToggle(!isOn); + } + + function handleDragEnd(e) { + if (isCloserToRightEdge(e)) { + updateToggle(true); + } else { + updateToggle(false); + } + } + + // ... +} +``` + +With this approach, both the `Toggle` component and its parent component update their state during the event. React [batches updates](/learn/queueing-a-series-of-state-updates) from different components together, so there will only be one render pass. + +You might also be able to remove the state altogether, and instead receive `isOn` from the parent component: + +```js {1,2} +// ✅ Also good: the component is fully controlled by its parent +function Toggle({ isOn, onChange }) { + function handleClick() { + onChange(!isOn); + } + + function handleDragEnd(e) { + if (isCloserToRightEdge(e)) { + onChange(true); + } else { + onChange(false); + } + } + + // ... +} +``` + +["Lifting state up"](/learn/sharing-state-between-components) lets the parent component fully control the `Toggle` by toggling the parent's own state. This means the parent component will have to contain more logic, but there will be less state overall to worry about. Whenever you try to keep two different state variables synchronized, try lifting state up instead! + +### Passing data to the parent {/*passing-data-to-the-parent*/} + +This `Child` component fetches some data and then passes it to the `Parent` component in an Effect: + +```js {9-14} +function Parent() { + const [data, setData] = useState(null); + // ... + return ; +} + +function Child({ onFetched }) { + const data = useSomeAPI(); + // 🔴 Avoid: Passing data to the parent in an Effect + useEffect(() => { + if (data) { + onFetched(data); + } + }, [onFetched, data]); + // ... +} +``` + +In React, data flows from the parent components to their children. When you see something wrong on the screen, you can trace where the information comes from by going up the component chain until you find which component passes the wrong prop or has the wrong state. When child components update the state of their parent components in Effects, the data flow becomes very difficult to trace. Since both the child and the parent need the same data, let the parent component fetch that data, and *pass it down* to the child instead: + +```js {4-5} +function Parent() { + const data = useSomeAPI(); + // ... + // ✅ Good: Passing data down to the child + return ; +} + +function Child({ data }) { + // ... +} +``` + +This is simpler and keeps the data flow predictable: the data flows down from the parent to the child. + +### Subscribing to an external store {/*subscribing-to-an-external-store*/} + +Sometimes, your components may need to subscribe to some data outside of the React state. This data could be from a third-party library or a built-in browser API. Since this data can change without React's knowledge, you need to manually subscribe your components to it. This is often done with an Effect, for example: + +```js {2-17} +function useOnlineStatus() { + // Not ideal: Manual store subscription in an Effect + const [isOnline, setIsOnline] = useState(true); + useEffect(() => { + function updateState() { + setIsOnline(navigator.onLine); + } + + updateState(); + + window.addEventListener('online', updateState); + window.addEventListener('offline', updateState); + return () => { + window.removeEventListener('online', updateState); + window.removeEventListener('offline', updateState); + }; + }, []); + return isOnline; +} + +function ChatIndicator() { + const isOnline = useOnlineStatus(); + // ... +} +``` + +Here, the component subscribes to an external data store (in this case, the browser `navigator.onLine` API). Since this API does not exist on the server (so it can't be used for the initial HTML), initially the state is set to `true`. Whenever the value of that data store changes in the browser, the component updates its state. + +Although it's common to use Effects for this, React has a purpose-built Hook for subscribing to an external store that is preferred instead. Delete the Effect and replace it with a call to [`useSyncExternalStore`](/reference/react/useSyncExternalStore): + +```js {11-16} +function subscribe(callback) { + window.addEventListener('online', callback); + window.addEventListener('offline', callback); + return () => { + window.removeEventListener('online', callback); + window.removeEventListener('offline', callback); + }; +} + +function useOnlineStatus() { + // ✅ Good: Subscribing to an external store with a built-in Hook + return useSyncExternalStore( + subscribe, // React won't resubscribe for as long as you pass the same function + () => navigator.onLine, // How to get the value on the client + () => true // How to get the value on the server + ); +} + +function ChatIndicator() { + const isOnline = useOnlineStatus(); + // ... +} +``` + +This approach is less error-prone than manually syncing mutable data to React state with an Effect. Typically, you'll write a custom Hook like `useOnlineStatus()` above so that you don't need to repeat this code in the individual components. [Read more about subscribing to external stores from React components.](/reference/react/useSyncExternalStore) + +### Fetching data {/*fetching-data*/} + +Many apps use Effects to kick off data fetching. It is quite common to write a data fetching Effect like this: + +```js {5-10} +function SearchResults({ query }) { + const [results, setResults] = useState([]); + const [page, setPage] = useState(1); + + useEffect(() => { + // 🔴 Avoid: Fetching without cleanup logic + fetchResults(query, page).then(json => { + setResults(json); + }); + }, [query, page]); + + function handleNextPageClick() { + setPage(page + 1); + } + // ... +} +``` + +You *don't* need to move this fetch to an event handler. + +This might seem like a contradiction with the earlier examples where you needed to put the logic into the event handlers! However, consider that it's not *the typing event* that's the main reason to fetch. Search inputs are often prepopulated from the URL, and the user might navigate Back and Forward without touching the input. + +It doesn't matter where `page` and `query` come from. While this component is visible, you want to keep `results` [synchronized](/learn/synchronizing-with-effects) with data from the network for the current `page` and `query`. This is why it's an Effect. + +However, the code above has a bug. Imagine you type `"hello"` fast. Then the `query` will change from `"h"`, to `"he"`, `"hel"`, `"hell"`, and `"hello"`. This will kick off separate fetches, but there is no guarantee about which order the responses will arrive in. For example, the `"hell"` response may arrive *after* the `"hello"` response. Since it will call `setResults()` last, you will be displaying the wrong search results. This is called a ["race condition"](https://en.wikipedia.org/wiki/Race_condition): two different requests "raced" against each other and came in a different order than you expected. + +**To fix the race condition, you need to [add a cleanup function](/learn/synchronizing-with-effects#fetching-data) to ignore stale responses:** + +```js {5,7,9,11-13} +function SearchResults({ query }) { + const [results, setResults] = useState([]); + const [page, setPage] = useState(1); + useEffect(() => { + let ignore = false; + fetchResults(query, page).then(json => { + if (!ignore) { + setResults(json); + } + }); + return () => { + ignore = true; + }; + }, [query, page]); + + function handleNextPageClick() { + setPage(page + 1); + } + // ... +} +``` + +This ensures that when your Effect fetches data, all responses except the last requested one will be ignored. + +Handling race conditions is not the only difficulty with implementing data fetching. You might also want to think about caching responses (so that the user can click Back and see the previous screen instantly), how to fetch data on the server (so that the initial server-rendered HTML contains the fetched content instead of a spinner), and how to avoid network waterfalls (so that a child can fetch data without waiting for every parent). + +**These issues apply to any UI library, not just React. Solving them is not trivial, which is why modern [frameworks](/learn/start-a-new-react-project#full-stack-frameworks) provide more efficient built-in data fetching mechanisms than fetching data in Effects.** + +If you don't use a framework (and don't want to build your own) but would like to make data fetching from Effects more ergonomic, consider extracting your fetching logic into a custom Hook like in this example: + +```js {4} +function SearchResults({ query }) { + const [page, setPage] = useState(1); + const params = new URLSearchParams({ query, page }); + const results = useData(`/api/search?${params}`); + + function handleNextPageClick() { + setPage(page + 1); + } + // ... +} + +function useData(url) { + const [data, setData] = useState(null); + useEffect(() => { + let ignore = false; + fetch(url) + .then(response => response.json()) + .then(json => { + if (!ignore) { + setData(json); + } + }); + return () => { + ignore = true; + }; + }, [url]); + return data; +} +``` + +You'll likely also want to add some logic for error handling and to track whether the content is loading. You can build a Hook like this yourself or use one of the many solutions already available in the React ecosystem. **Although this alone won't be as efficient as using a framework's built-in data fetching mechanism, moving the data fetching logic into a custom Hook will make it easier to adopt an efficient data fetching strategy later.** + +In general, whenever you have to resort to writing Effects, keep an eye out for when you can extract a piece of functionality into a custom Hook with a more declarative and purpose-built API like `useData` above. The fewer raw `useEffect` calls you have in your components, the easier you will find to maintain your application. + + + +- If you can calculate something during render, you don't need an Effect. +- To cache expensive calculations, add `useMemo` instead of `useEffect`. +- To reset the state of an entire component tree, pass a different `key` to it. +- To reset a particular bit of state in response to a prop change, set it during rendering. +- Code that runs because a component was *displayed* should be in Effects, the rest should be in events. +- If you need to update the state of several components, it's better to do it during a single event. +- Whenever you try to synchronize state variables in different components, consider lifting state up. +- You can fetch data with Effects, but you need to implement cleanup to avoid race conditions. + + + + + +#### Transform data without Effects {/*transform-data-without-effects*/} + +The `TodoList` below displays a list of todos. When the "Show only active todos" checkbox is ticked, completed todos are not displayed in the list. Regardless of which todos are visible, the footer displays the count of todos that are not yet completed. + +Simplify this component by removing all the unnecessary state and Effects. + + + +```js +import { useState, useEffect } from 'react'; +import { initialTodos, createTodo } from './todos.js'; + +export default function TodoList() { + const [todos, setTodos] = useState(initialTodos); + const [showActive, setShowActive] = useState(false); + const [activeTodos, setActiveTodos] = useState([]); + const [visibleTodos, setVisibleTodos] = useState([]); + const [footer, setFooter] = useState(null); + + useEffect(() => { + setActiveTodos(todos.filter(todo => !todo.completed)); + }, [todos]); + + useEffect(() => { + setVisibleTodos(showActive ? activeTodos : todos); + }, [showActive, todos, activeTodos]); + + useEffect(() => { + setFooter( +
+ {activeTodos.length} todos left +
+ ); + }, [activeTodos]); + + return ( + <> + + setTodos([...todos, newTodo])} /> +
    + {visibleTodos.map(todo => ( +
  • + {todo.completed ? {todo.text} : todo.text} +
  • + ))} +
+ {footer} + + ); +} + +function NewTodo({ onAdd }) { + const [text, setText] = useState(''); + + function handleAddClick() { + setText(''); + onAdd(createTodo(text)); + } + + return ( + <> + setText(e.target.value)} /> + + + ); +} +``` + +```js src/todos.js +let nextId = 0; + +export function createTodo(text, completed = false) { + return { + id: nextId++, + text, + completed + }; +} + +export const initialTodos = [ + createTodo('Get apples', true), + createTodo('Get oranges', true), + createTodo('Get carrots'), +]; +``` + +```css +label { display: block; } +input { margin-top: 10px; } +``` + +
+ + + +If you can calculate something during rendering, you don't need state or an Effect that updates it. + + + + + +There are only two essential pieces of state in this example: the list of `todos` and the `showActive` state variable which represents whether the checkbox is ticked. All of the other state variables are [redundant](/learn/choosing-the-state-structure#avoid-redundant-state) and can be calculated during rendering instead. This includes the `footer` which you can move directly into the surrounding JSX. + +Your result should end up looking like this: + + + +```js +import { useState } from 'react'; +import { initialTodos, createTodo } from './todos.js'; + +export default function TodoList() { + const [todos, setTodos] = useState(initialTodos); + const [showActive, setShowActive] = useState(false); + const activeTodos = todos.filter(todo => !todo.completed); + const visibleTodos = showActive ? activeTodos : todos; + + return ( + <> + + setTodos([...todos, newTodo])} /> +
    + {visibleTodos.map(todo => ( +
  • + {todo.completed ? {todo.text} : todo.text} +
  • + ))} +
+
+ {activeTodos.length} todos left +
+ + ); +} + +function NewTodo({ onAdd }) { + const [text, setText] = useState(''); + + function handleAddClick() { + setText(''); + onAdd(createTodo(text)); + } + + return ( + <> + setText(e.target.value)} /> + + + ); +} +``` + +```js src/todos.js +let nextId = 0; + +export function createTodo(text, completed = false) { + return { + id: nextId++, + text, + completed + }; +} + +export const initialTodos = [ + createTodo('Get apples', true), + createTodo('Get oranges', true), + createTodo('Get carrots'), +]; +``` + +```css +label { display: block; } +input { margin-top: 10px; } +``` + +
+ +
+ +#### Cache a calculation without Effects {/*cache-a-calculation-without-effects*/} + +In this example, filtering the todos was extracted into a separate function called `getVisibleTodos()`. This function contains a `console.log()` call inside of it which helps you notice when it's being called. Toggle "Show only active todos" and notice that it causes `getVisibleTodos()` to re-run. This is expected because visible todos change when you toggle which ones to display. + +Your task is to remove the Effect that recomputes the `visibleTodos` list in the `TodoList` component. However, you need to make sure that `getVisibleTodos()` does *not* re-run (and so does not print any logs) when you type into the input. + + + +One solution is to add a `useMemo` call to cache the visible todos. There is also another, less obvious solution. + + + + + +```js +import { useState, useEffect } from 'react'; +import { initialTodos, createTodo, getVisibleTodos } from './todos.js'; + +export default function TodoList() { + const [todos, setTodos] = useState(initialTodos); + const [showActive, setShowActive] = useState(false); + const [text, setText] = useState(''); + const [visibleTodos, setVisibleTodos] = useState([]); + + useEffect(() => { + setVisibleTodos(getVisibleTodos(todos, showActive)); + }, [todos, showActive]); + + function handleAddClick() { + setText(''); + setTodos([...todos, createTodo(text)]); + } + + return ( + <> + + setText(e.target.value)} /> + +
    + {visibleTodos.map(todo => ( +
  • + {todo.completed ? {todo.text} : todo.text} +
  • + ))} +
+ + ); +} +``` + +```js src/todos.js +let nextId = 0; +let calls = 0; + +export function getVisibleTodos(todos, showActive) { + console.log(`getVisibleTodos() was called ${++calls} times`); + const activeTodos = todos.filter(todo => !todo.completed); + const visibleTodos = showActive ? activeTodos : todos; + return visibleTodos; +} + +export function createTodo(text, completed = false) { + return { + id: nextId++, + text, + completed + }; +} + +export const initialTodos = [ + createTodo('Get apples', true), + createTodo('Get oranges', true), + createTodo('Get carrots'), +]; +``` + +```css +label { display: block; } +input { margin-top: 10px; } +``` + +
+ + + +Remove the state variable and the Effect, and instead add a `useMemo` call to cache the result of calling `getVisibleTodos()`: + + + +```js +import { useState, useMemo } from 'react'; +import { initialTodos, createTodo, getVisibleTodos } from './todos.js'; + +export default function TodoList() { + const [todos, setTodos] = useState(initialTodos); + const [showActive, setShowActive] = useState(false); + const [text, setText] = useState(''); + const visibleTodos = useMemo( + () => getVisibleTodos(todos, showActive), + [todos, showActive] + ); + + function handleAddClick() { + setText(''); + setTodos([...todos, createTodo(text)]); + } + + return ( + <> + + setText(e.target.value)} /> + +
    + {visibleTodos.map(todo => ( +
  • + {todo.completed ? {todo.text} : todo.text} +
  • + ))} +
+ + ); +} +``` + +```js src/todos.js +let nextId = 0; +let calls = 0; + +export function getVisibleTodos(todos, showActive) { + console.log(`getVisibleTodos() was called ${++calls} times`); + const activeTodos = todos.filter(todo => !todo.completed); + const visibleTodos = showActive ? activeTodos : todos; + return visibleTodos; +} + +export function createTodo(text, completed = false) { + return { + id: nextId++, + text, + completed + }; +} + +export const initialTodos = [ + createTodo('Get apples', true), + createTodo('Get oranges', true), + createTodo('Get carrots'), +]; +``` + +```css +label { display: block; } +input { margin-top: 10px; } +``` + +
+ +With this change, `getVisibleTodos()` will be called only if `todos` or `showActive` change. Typing into the input only changes the `text` state variable, so it does not trigger a call to `getVisibleTodos()`. + +There is also another solution which does not need `useMemo`. Since the `text` state variable can't possibly affect the list of todos, you can extract the `NewTodo` form into a separate component, and move the `text` state variable inside of it: + + + +```js +import { useState, useMemo } from 'react'; +import { initialTodos, createTodo, getVisibleTodos } from './todos.js'; + +export default function TodoList() { + const [todos, setTodos] = useState(initialTodos); + const [showActive, setShowActive] = useState(false); + const visibleTodos = getVisibleTodos(todos, showActive); + + return ( + <> + + setTodos([...todos, newTodo])} /> +
    + {visibleTodos.map(todo => ( +
  • + {todo.completed ? {todo.text} : todo.text} +
  • + ))} +
+ + ); +} + +function NewTodo({ onAdd }) { + const [text, setText] = useState(''); + + function handleAddClick() { + setText(''); + onAdd(createTodo(text)); + } + + return ( + <> + setText(e.target.value)} /> + + + ); +} +``` + +```js src/todos.js +let nextId = 0; +let calls = 0; + +export function getVisibleTodos(todos, showActive) { + console.log(`getVisibleTodos() was called ${++calls} times`); + const activeTodos = todos.filter(todo => !todo.completed); + const visibleTodos = showActive ? activeTodos : todos; + return visibleTodos; +} + +export function createTodo(text, completed = false) { + return { + id: nextId++, + text, + completed + }; +} + +export const initialTodos = [ + createTodo('Get apples', true), + createTodo('Get oranges', true), + createTodo('Get carrots'), +]; +``` + +```css +label { display: block; } +input { margin-top: 10px; } +``` + +
+ +This approach satisfies the requirements too. When you type into the input, only the `text` state variable updates. Since the `text` state variable is in the child `NewTodo` component, the parent `TodoList` component won't get re-rendered. This is why `getVisibleTodos()` doesn't get called when you type. (It would still be called if the `TodoList` re-renders for another reason.) + +
+ +#### Reset state without Effects {/*reset-state-without-effects*/} + +This `EditContact` component receives a contact object shaped like `{ id, name, email }` as the `savedContact` prop. Try editing the name and email input fields. When you press Save, the contact's button above the form updates to the edited name. When you press Reset, any pending changes in the form are discarded. Play around with this UI to get a feel for it. + +When you select a contact with the buttons at the top, the form resets to reflect that contact's details. This is done with an Effect inside `EditContact.js`. Remove this Effect. Find another way to reset the form when `savedContact.id` changes. + + + +```js src/App.js hidden +import { useState } from 'react'; +import ContactList from './ContactList.js'; +import EditContact from './EditContact.js'; + +export default function ContactManager() { + const [ + contacts, + setContacts + ] = useState(initialContacts); + const [ + selectedId, + setSelectedId + ] = useState(0); + const selectedContact = contacts.find(c => + c.id === selectedId + ); + + function handleSave(updatedData) { + const nextContacts = contacts.map(c => { + if (c.id === updatedData.id) { + return updatedData; + } else { + return c; + } + }); + setContacts(nextContacts); + } + + return ( +
+ setSelectedId(id)} + /> +
+ +
+ ) +} + +const initialContacts = [ + { id: 0, name: 'Taylor', email: 'taylor@mail.com' }, + { id: 1, name: 'Alice', email: 'alice@mail.com' }, + { id: 2, name: 'Bob', email: 'bob@mail.com' } +]; +``` + +```js src/ContactList.js hidden +export default function ContactList({ + contacts, + selectedId, + onSelect +}) { + return ( +
+
    + {contacts.map(contact => +
  • + +
  • + )} +
+
+ ); +} +``` + +```js src/EditContact.js active +import { useState, useEffect } from 'react'; + +export default function EditContact({ savedContact, onSave }) { + const [name, setName] = useState(savedContact.name); + const [email, setEmail] = useState(savedContact.email); + + useEffect(() => { + setName(savedContact.name); + setEmail(savedContact.email); + }, [savedContact]); + + return ( +
+ + + + +
+ ); +} +``` + +```css +ul, li { + list-style: none; + margin: 0; + padding: 0; +} +li { display: inline-block; } +li button { + padding: 10px; +} +label { + display: block; + margin: 10px 0; +} +button { + margin-right: 10px; + margin-bottom: 10px; +} +``` + +
+ + + +It would be nice if there was a way to tell React that when `savedContact.id` is different, the `EditContact` form is conceptually a _different contact's form_ and should not preserve state. Do you recall any such way? + + + + + +Split the `EditContact` component in two. Move all the form state into the inner `EditForm` component. Export the outer `EditContact` component, and make it pass `savedContact.id` as the `key` to the inner `EditForm` component. As a result, the inner `EditForm` component resets all of the form state and recreates the DOM whenever you select a different contact. + + + +```js src/App.js hidden +import { useState } from 'react'; +import ContactList from './ContactList.js'; +import EditContact from './EditContact.js'; + +export default function ContactManager() { + const [ + contacts, + setContacts + ] = useState(initialContacts); + const [ + selectedId, + setSelectedId + ] = useState(0); + const selectedContact = contacts.find(c => + c.id === selectedId + ); + + function handleSave(updatedData) { + const nextContacts = contacts.map(c => { + if (c.id === updatedData.id) { + return updatedData; + } else { + return c; + } + }); + setContacts(nextContacts); + } + + return ( +
+ setSelectedId(id)} + /> +
+ +
+ ) +} + +const initialContacts = [ + { id: 0, name: 'Taylor', email: 'taylor@mail.com' }, + { id: 1, name: 'Alice', email: 'alice@mail.com' }, + { id: 2, name: 'Bob', email: 'bob@mail.com' } +]; +``` + +```js src/ContactList.js hidden +export default function ContactList({ + contacts, + selectedId, + onSelect +}) { + return ( +
+
    + {contacts.map(contact => +
  • + +
  • + )} +
+
+ ); +} +``` + +```js src/EditContact.js active +import { useState } from 'react'; + +export default function EditContact(props) { + return ( + + ); +} + +function EditForm({ savedContact, onSave }) { + const [name, setName] = useState(savedContact.name); + const [email, setEmail] = useState(savedContact.email); + + return ( +
+ + + + +
+ ); +} +``` + +```css +ul, li { + list-style: none; + margin: 0; + padding: 0; +} +li { display: inline-block; } +li button { + padding: 10px; +} +label { + display: block; + margin: 10px 0; +} +button { + margin-right: 10px; + margin-bottom: 10px; +} +``` + +
+ +
+ +#### Submit a form without Effects {/*submit-a-form-without-effects*/} + +This `Form` component lets you send a message to a friend. When you submit the form, the `showForm` state variable is set to `false`. This triggers an Effect calling `sendMessage(message)`, which sends the message (you can see it in the console). After the message is sent, you see a "Thank you" dialog with an "Open chat" button that lets you get back to the form. + +Your app's users are sending way too many messages. To make chatting a little bit more difficult, you've decided to show the "Thank you" dialog *first* rather than the form. Change the `showForm` state variable to initialize to `false` instead of `true`. As soon as you make that change, the console will show that an empty message was sent. Something in this logic is wrong! + +What's the root cause of this problem? And how can you fix it? + + + +Should the message be sent _because_ the user saw the "Thank you" dialog? Or is it the other way around? + + + + + +```js +import { useState, useEffect } from 'react'; + +export default function Form() { + const [showForm, setShowForm] = useState(true); + const [message, setMessage] = useState(''); + + useEffect(() => { + if (!showForm) { + sendMessage(message); + } + }, [showForm, message]); + + function handleSubmit(e) { + e.preventDefault(); + setShowForm(false); + } + + if (!showForm) { + return ( + <> +

Thanks for using our services!

+ + + ); + } + + return ( +
+