Let's say I have a np.array such as:
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10
Given a window of size 5, I want to and get:
0, 1, 2, 3, 4
1, 2, 3, 4, 5
2, 3, 4, 5, 6
3, 4, 5, 6, 7
4, 5, 6, 7, 8
5, 6, 7, 8, 9
6, 7, 8, 9, 10
It's an easy problem any rookie can solve, just wanted to know what is the most Pythonic numpy implementation.
Depends on how badly you want to avoid copying that data. If badly enough, use np.lib.stride_tricks.as_strided:
arr = np.arange(11)
window = 7
as_strided(arr, shape=(window, arr.size - window + 1), strides=arr.strides * 2, writeable=False)
This will make an array that has views of the same element in multiple places, so I've set writable=False. You can always make an expanded copy if you like.
