Numpy difference between neighboring elements

I have algorithm of calculation of the difference between neighboring elements in pure python:

    a = range(1000000) #it's numpy array in my case
    prev = a[0]
    b = [0, ]
    for i in a[1:]:
        b.append(i - prev)
        prev = i

Is there any way to rewrite this functions with Numpy?

Solution 1:

There is the diff method:

a = range(5) # python list of numpy array
np.diff(a)

returns

array([1, 1, 1, 1])

Solution 2:

Because he said

#it's numpy array in my case,

you may want to use a[1:]-a[:-1] as it's faster.

> %%timeit a=np.random.rand(10000)
> a[1:]-a[:-1]
6.94 µs ± 188 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)

> %%timeit a=np.random.rand(10000)
> np.diff(a)
15.7 µs ± 247 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)