Search code examples
numpytest-suite

numpy: test that the array is almost 0

To check that the array is almost 0, I use np.testing.assert_almost_equal:

np.testing.assert_almost_equal(v, np.zeros(len(v)))

This allocates a new vector just for the test.

Is there a better way?

Solution

  • You don't need to pass an array, a scalar is enough and numpy will broadcast:

    v = np.array([0, 0.0000001])

np.testing.assert_almost_equal(v, 0)

    As a bonus, this works with any number of dimensions.