An Enzyme Gotcha

Enzyme is a library for testing React components. It’s from the good folks at AirBnB, who already provide a great set of linter defaults.

Enzyme’s selling point is a jQuery like syntax for finding elements to assert against. Compare:


with React’s TestUtils:

TestUtils.scryRenderedDOMComponentsWithTag(component, 'a');  

And unlike TestUtils, Enzyme works with shallow rendering. They even simplify the process:

import { shallow } from 'enzyme';

const wrapper = shallow(<My Component />);  

But the jQuery-like syntax can trip you up.

Suppose you wanted simulate clicking a link with the text ‘Click Me!’. You might try something like this:

    .findWhere(n => n.text() === 'Click Me!');

But that fails with Enzyme saying:

This method is only meant to be run on single node. 2 found instead.  

What happened?

Unlike jQuery, Enzyme doesn’t return elements, it returns wrapped component trees.

Thus, our first call, component.find(‘a’) returns an array of wrapped a component trees. Assuming there’s a single link on the page, component.find(‘a’) returns an array containing a single tree.

Next we call findWhere(n=> n.text() === 'Click Me!') This iterates over the single a component tree, which includes two nodes: the a itself and the text node containing Click Me!.

Calling .text() on either node will return Click Me! and thus the call returns two nodes. This isn’t what we want.

Consider this instead:

component.findWhere(n => n.type() === 'a' && n.contains('Click Me!'))  

Spot the difference? The above restricts the results to nodes of type a. This eliminates the text nodes, and ensures we get our single link tree.

Given Enzyme’s rapid evolution, I wouldn’t be surprised to see this incorporated as a helper method, maybe something like findWithTypeAndText(‘a’, 'Click Me!'), but until then, the above will do.