React conditional className, empty strings and null

When developing React components, you often need to conditionally apply a className attribute to one or more elements. Sometimes, you will have two or more possible values depending on a condition. But there are also times that you might apply a className based on a condition or leave it completely empty otherwise.

There is a correct way to handle a conditional empty className and an incorrect one. Surprisingly, the incorrect way is pretty common and examples of it can be found all around the web. Consider the following code:

const MyComponent = ({ enabled }) => {
  return ( <div className={enabled ? 'enabled' : ''}> Hi </div> );
};

const OtherComponent = ({ enabled }) => {
  return ( <div className={enabled ? 'enabled' : null}> Hi </div> );
};

ReactDOM.createRoot(document.getElementById('root')).render(
  <>
    <MyComponent enabled={false} />
    <OtherComponent enabled={false} />
  </>
);

In this code example, we define two very similar components. Both of them conditionally set the className of an element based on the value of the enabled prop. The first one will set the className to an empty string if enabled is false and the second one will set it to null.

The resulting output is pretty similar. However, if you carefully inspect the HTML, you will notice that the first one will render <div class>Hi</div> whereas the second one will render <div>Hi</div>. This kind of markup (an attribute being present but without value) is rather uncommon and you'd rarely ever see something like that without React. This subtle difference is quite important and might be the root of a lot of problems, especially when writing CSS selectors for elements with/without any classes (e.g. [class]/:not([class])).

Therefore, you should prefer null when you don't want to add a className to an element, instead of an empty string. It keeps the markup cleaner and might help prevent some potential issues.