Use of index to get sample value in knn classifier

Question

As per documentation:

As you can see, it returns [[0.5]], and [[2]], which means that the element is at distance 0.5 and is the third element of samples (ind