numpy.where的Output

Question

I noticed a strange thing concerning the output of numpy.where. According to the documentation the output should be of type ndarray. If I provide all arguments (including also the optional ones) this is true, eg

np.where(foo == 2, foo, -1)
>>> array([-1,  2, -1, -1])

However, if I insert an argument like

np.where(foo == 2)
>>> (array([1], dtype=int64),)

I get an array encapsulated in a tuple. Does anyone know why that would make any sense?

Answer 1

The answer is in the documentation you provided :

When only condition is provided, this function is a shorthand for np.asarray(condition).nonzero() .

Looking at np.nonzero() 's documentation , we can see that it returns a tuple of arrays.