I am new to react Hooks. Am trying to make use of useState in my code. While I was using it I found a term "Lazy initial state"
https://reactjs.org/docs/hooks-reference.html#lazy-initial-state
const [state, setState] = useState(() => {
const initialState = someExpensiveComputation(props);
return initialState;
});
But I am not able to think of any use case where this lazy initialising of state will be useful.
Like say my DOM is rendering and it needs the state value, but my useState has not initialised it yet! And say if you have rendered the DOM and the someExpensiveComputation has finished, the DOM will re-render!
The value passed to the useState hook in the initial render is the initial state value, and gets disregarded in subsequent renders. This initial value can be the result of calling a function as in the following situation:
const Component = () => {
const [state, setState] = useState(getInitialHundredItems())
}
But note that getInitialHundredItems is unconditionally and needlessly called on each render cycle.
For use cases like this instead of just calling a function that returns a value you can pass a function which returns the initial state. This function will only be executed once (initial render) and not on each render like the above code will. See Lazy Initial State for details.
const Component = () =>{
const [state, setState] = useState(getInitialHundredItems)
}
Related
I am new to react Hooks. Am trying to make use of useState in my code. While I was using it I found a term "Lazy initial state"
https://reactjs.org/docs/hooks-reference.html#lazy-initial-state
const [state, setState] = useState(() => {
const initialState = someExpensiveComputation(props);
return initialState;
});
But I am not able to think of any use case where this lazy initialising of state will be useful.
Like say my DOM is rendering and it needs the state value, but my useState has not initialised it yet! And say if you have rendered the DOM and the someExpensiveComputation has finished, the DOM will re-render!
The value passed to the useState hook in the initial render is the initial state value, and gets disregarded in subsequent renders. This initial value can be the result of calling a function as in the following situation:
const Component = () => {
const [state, setState] = useState(getInitialHundredItems())
}
But note that getInitialHundredItems is unconditionally and needlessly called on each render cycle.
For use cases like this instead of just calling a function that returns a value you can pass a function which returns the initial state. This function will only be executed once (initial render) and not on each render like the above code will. See Lazy Initial State for details.
const Component = () =>{
const [state, setState] = useState(getInitialHundredItems)
}
I saw React codes generally use arrow function for setState(). However, for useState(), I see that they just put a number, eg useState(0) instead of arrow function, ie useState( () => 0 )
Are there any reasons why I would want to use an arrow function in useState ? Is it alright to use only arrow functions for ALL React hook regardless, including initialising the hook. Any gotcha ? Thanks very much.
For your simple case it does not matter. You are just making an extra function call which returns a value without any computation.
But in case your initial state is a result of an expensive computation you can use a function.
Using the below
const [state, setState] = useState(() => {
const initialState = someExpensiveComputation(props);
return initialState;
});
instead of
const [state, setState] = useState(someExpensiveComputation(props));
matters greatly.
The initialState argument is the state used during the initial render. In subsequent renders, it is disregarded.
So even when your value is disregarded you would be making that expensive computation. You can provide a function in your useState which will only run on the very first render, like in the second case above. [Link](The initialState argument is the state used during the initial render. In subsequent renders, it is disregarded.)
As far as I know, the arrow functions in state setter permit use the previous state.
For example:
const [message, setMessage] = useState('hello ')
... # Receive data from a form or API, etc.
let name = 'Richard'
setMessage(old_state => old_state + name) # 'hello Richard'
This way, you can create counter or reuse some data of the previous state.
Use setState(new_vale) is totally correct for reset the value.
What is the case where you use useState's initial value as a function?
Is there any difference from just passing a value?
e.g.
const [state, setState] = useState(() => someValue)
You use it when you want the computation of that initial state to happen only once. Because if you use an expression instead say:
const [state, setState] = useState(compute())
The compute runs on other renders too, just its value is ignored after first* render.
So if you do:
const [state, setState] = useState(() => compute())
Now, compute will run only once.
From the docs:
const [state, setState] = useState(initialState);
The initialState argument is the state used during the initial render.
In subsequent renders, it is disregarded. If the initial state is the
result of an expensive computation, you may provide a function
instead, which will be executed only on the initial render
const [state, setState] = useState(() => {
const initialState = someExpensiveComputation(props);
return initialState;
});
* Well if it is strict mode then it could be the value of first render gets ignored too due to double invoking the render method. But this is not important for this answer. Because the value would now be ignored after second render.
If you want to use useState's initial value as a function, you need to use currying :
const [state, setState] = useState(() => () => someValue);
This is because in the documentation, useState executes the provided function and considers its result as the initial value.
Using currying, () => someValue is returned and considered to be the intial value.
When we have some heavy computation to initialize the state, we should use the function. And this is lazy initialization of the state. Here is a well-written blog on react state lazy initialization by kentcdodds.
The reason why wrapping in a function is less computationally heavy is because react will rerun the component function when props change. If useState() has a computation inside like useState(heavyComputation()), JavaScript will run heavyComputation(), but in useState(()=>heavyComputation()), JavaScript will not run heavyComputation(), but will pass in the curried function, and useState knows not to rerun this function.
Normal
Render 1: useState(heavyComputation()) //JavaScript calls heavyComputation
Render 2: useState(heavyComputation()) //JavaScript calls heavyComputation
Render 3: useState(heavyComputation()) //JavaScript calls heavyComputation
...
Wrapped in Function (curried)
Render 1: useState(()=>heavyComputation()) //useState() calls heavyComputation
Render 2: useState(()=>heavyComputation()) //useState() uses the value from the previous render so heavyComputation() is not called
Render 3: useState(()=>heavyComputation()) //useState() uses the value from the previous render so heavyComputation() is not called
...
I am looking for a way to perform more advanced comparison instead of the second parameter of the useEffect React hook.
Specifically, I am looking for something more like this:
useEffect(
() => doSomething(),
[myInstance],
(prev, curr) => { /* compare prev[0].value with curr[0].value */ }
);
Is there anything I missed from the React docs about this or is there any way of implementing such a hook on top of what already exists, please?
If there is a way to implement this, this is how it would work: the second parameter is an array of dependencies, just like the useEffect hook coming from React, and the third is a callback with two parameters: the array of dependencies at the previous render, and the array of dependencies at the current render.
you could use React.memo function:
const areEqual = (prevProps, nextProps) => {
return (prevProps.title === nextProps.title)
};
export default React.memo(Component, areEqual);
or use custom hooks for that:
How to compare oldValues and newValues on React Hooks useEffect?
In class based components was easy to perform a deep comparison. componentDidUpdate provides a snapshot of previous props and previous state
componentDidUpdate(prevProps, prevState, snapshot){
if(prevProps.foo !== props.foo){ /* ... */ }
}
The problem is useEffect it's not exactly like componentDidUpdate. Consider the following
useEffect(() =>{
/* action() */
},[props])
The only thing you can assert about the current props when action() gets called is that it changed (shallow comparison asserts to false). You cannot have a snapshot of prevProps cause hooks are like regular functions, there aren't part of a lifecycle (and don't have an instance) which ensures synchronicity (and inject arguments). Actually the only thing ensuring hooks value integrity is the order of execution.
Alternatives to usePrevious
Before updating check if the values are equal
const Component = props =>{
const [foo, setFoo] = useState('bar')
const updateFoo = val => foo === val ? null : setFoo(val)
}
This can be useful in some situations when you need to ensure an effect to run only once(not useful in your use case).
useMemo:
If you want to perform comparison to prevent unecessary render calls (shoudComponentUpdate), then useMemo is the way to go
export React.useMemo(Component, (prev, next) => true)
But when you need to get access to the previous value inside an already running effect there is no alternatives left. Cause if you already are inside useEffect it means that the dependency it's already updated (current render).
Why usePrevious works
useRef isn't just for refs, it's a very straightforward way to mutate values without triggering a re render. So the cycle is the following
Component gets mounted
usePrevious stores the inital value inside current
props changes triggering a re render inside Component
useEffect is called
usePrevious is called
Notice that the usePrevious is always called after the useEffect (remember, order matters!). So everytime you are inside an useEffect the current value of useRef will always be one render behind.
const usePrevious = value =>{
const ref = useRef()
useEffect(() => ref.current = value,[value])
}
const Component = props =>{
const { A } = props
useEffect(() =>{
console.log('runs first')
},[A])
//updates after the effect to store the current value (which will be the previous on next render
const previous = usePrevious(props)
}
I hit the same problem recently and a solution that worked for me is to create a custom useStateWithCustomComparator hook.
In your the case of your example that would mean to replace
const [myInstance, setMyInstance] = useState(..)
with
const [myInstance, setMyInstance] = useStateWithCustomComparator(..)
The code for my custom hook in Typescript looks like that:
const useStateWithCustomComparator = <T>(initialState: T, customEqualsComparator: (obj1: T, obj2: T) => boolean) => {
const [state, setState] = useState(initialState);
const changeStateIfNotEqual = (newState: any) => {
if (!customEqualsComparator(state, newState)) {
setState(newState);
}
};
return [state, changeStateIfNotEqual] as const;
};
I am new to hooks so I will ask this question:
How can I legally use setState inside a onChange function ??
{
const state = useState([{date: date}]);
};
This throws an error :(
Expected: update the state
And result error message: Hooks can only be called inside the body of a function component.
I did find some answers on the web about multiple React version installed but this is not the case here :)
You need to declare your hook in the body of your main component function:
const MyComponent = () => {
// declaring your useState Hook - it returns a getter and setter
const [date, setDate] = useState(null)
const myCallback = (newDate)=>{
// you can read the state
const state = date;
// or, you can do the equivalent of setState
setDate(newDate)
}
return // return your .jsx
}
The things to note are that you declare your hook once, and that it returns two parameters, a getter and a setter. From then on, you interact with that original Hook using the getter and setter only.
For your pastebin code:
const [myState,setMyState] = useState(
{
name: "",
type: props.navigation.getParam("serviceName"),
date: "15-05-2018",
imageURI: props.navigation.getParam("imageURI")
});
const handleChange = e => {
setMyState(prevState=> {...prevState, name: e.nativeEvent.text})
};
(I've used myState and setMyState but they can be whatever you like)
Hmm it seems you've got it wrong...
There's no single state when we use hooks like in the case of a class,
const [item, setValue] = useState(null)
this sets the state variable "item" with initial value of null. Now we can use the "setValue" to change the value of "item" to what we want like this
setValue(5).There is no setState as in the case of a class component. Since the value of a state variable has been changed the component rerenders. We can declare multiple state-variables like
const [newitem, setnewItemValue] = useState(null) and whenever a state-variable changes the component will rerender