numpy: ndarray index with repeated values [duplicate]

Consider I have an array:

arr2 = np.array([10., 10., 10.,  10.,  10.,  10., 10., 10., 10., 10.])

and an index with repeated values:

idx = np.array([3,3,3,4,4,5])

If I do so:

arr2[idx] -=1

then I got this:

array([10., 10., 10.,  9.,  9.,  9., 10., 10., 10., 10.])

But I want to subtract by each value in index, taking into account the repetitions! How to get this?

array([10., 10., 10.,  7.,  8.,  9., 10., 10., 10., 10.])