a = np.ramdon.rand(10)
# We can index a numpy array by [start:stop(:step)]
a[::1]  # returns [a[0], a[1], ..., a[9]]
a[::-1]  # returns [a[9], a[8], ..., a[0]]
As shown in the above example, start and stop for both a[::1] and a[::-1] should be by default, since only the step arguments are given. For a[::1], I suppose the default start should be 0 and the default stop should be -1 (i.e., a[::1] is the same as a[0:-1:1]), but apparently this is not true for a[::-1], since a[0:-1:-1] returns an empty array. So what are the exact default setting for stop and start? I am so confused because all the documentations I ran into only mentioned that to reverse an array you can use ::-1 with no further explanation.
