I understand that the special 'key' prop when used with child components that are created dynamically from arrays helps React to uniquely identify components and render updates efficiently. But I would like to know when and why would it be necessary to use the key prop for a 'non-dynamic' component.
My application uses a Reducer and useContext hook to manage state for a Functional Component A. The state object has a maximum 3 levels of nesting. Component A updates state and passes part of the state object as props to two instances of a child component B. B uses these props to render a switch component and 2 input components. Here's the simplified code for this hierarchy.
Component A:
const A: FC = () => {
// ....
// graphql queries to get data and update state using reducer
// ...
return (
<B
enabled={data.a.b.enabled}
value1={data.a.b.value1}
value2={data.a.b.value2}
/>
<B
enabled={data.a.b.enabled}
value1={data.a.b.value1}
value2={data.a.b.value2}
/>
);
};
Component B:
const B: FC = props => {
const { value1, value2, enabled} = props; // there are other props as well
return (
<>
<div className={someClassLogic}>
<Switch
onChange={onValueChange}
isChecked={enabled}
disabled={disabled}
/>
</div>
<div className={someClassLogic} >
<Input
input={value1}
disabled={disabled}
/>
</div>
<div className={someClassLogic}>
<Input
input={value2}
disabled={disabled}
/>
</div>
</>
);
};
A tablerow click event is used to update the state and the component B displays the 'settings' of this selected item which the user can mutate using the component B.
Here's the problem I'm facing- when the state is updated by a user action (selecting a row from a table, not shown in the snippet), I can see that both A and B receive the new data in the react developer tools and by printing to the console. But, a render does not take place to show the new data. I would like to understand why this is the case.
After looking up this issue, I figured I need a key prop while instantiating component B (the answers don't clearly explain why). With the following addition, the values did render correctly. Why is a key necessary here and why does it only work when the key contains all props that can change values? If I only use the uniqueId as the key, the value1 and value2 do not render correctly again. If I have many changing-props, do I have to them to add the key as well? Isn't there a less clumsy approach?
Updated Component A:
const A: FC = () => {
return (
<B
key={`${data.a.uniqueId}-
${data.a.b.value1}-
${data.a.b.value2}
enabled={data.a.b.enabled}
value1={data.a.b.value1}
value2={data.a.b.value2}
/>
<B
key={`${data.a.uniqueId}-
${data.a.b.value1}-
${data.a.b.value2}
enabled={data.a.b.enabled}
value1={data.a.b.value1}
value2={data.a.b.value2}
/>
);
};
Also, I noticed that although clicking on a table row rendered the correct value in component B now, but clicking on a row that is not changed by the user so far would cause the previously rendered values to remain on the Input1 and Input2 components (instead of blank). So I had to add keys to the Inputs as well with the enabled state attached to it which fixed this issue.
Updated Component B:
const B: FC = props => {
const { value1, value2, enabled} = props; // there are other props as well
return (
<>
<div className={someClassLogic}>
<Switch
onChange={onValueChange}
isChecked={enabled}
disabled={disabled}
/>
</div>
<div className={someClassLogic} >
<Input
key={`value1-${enabled}`}
input={value1}
disabled={disabled}
/>
</div>
<div className={someClassLogic}>
<Input
key={`value2-${enabled}`}
input={value2}
disabled={disabled}
/>
</div>
</>
);
};
Here again, why is a key needed? Doesn't react figure out that the props have changed and automatically render again?