k=[False, False,  True],
    fill_value=999999)

When `indices` is not one-dimensional, the output also has these dimensions:

>>> np.ma.take(a, [[0, 1], [2, 3]])
masked_array(data=[[4, 3],
                [--, 7]],
            mask=[[False, False],
                [ True, False]],
    fill_value=999999)
r