How to shift the entire numpy array, with wrapping

Question

Not sure how to best title this question, but basically I would like to generate a new numpy array to be based on an existing array. The only difference is that the values have been shifted to an index I specify. Also assume that wrapping is required.

For simplicity, consider the base array:

[[0,1,2],
 [3,4,5],
 [6,7,8]]

If I want the zero (0) or the first element from the base array to be shifted at (0,1), it will be:

[[2,0,1],
 [5,3,4],
 [8,6,7]]

If I want the first element moved at (2,2) it will be:

[[4,5,3],
 [7,8,6],
 [1,2,0]]

I assume the 5 in matrix_2 position [2, 0] should be an 8. Can't edit because the edit is "too short" per SO requirements. — ShlomiF
– ShlomiF, Commented Nov 26, 2018 at 6:37
Depending on how often you do this, you might want to consider a ring buffer instead. The answer recommending np.roll is O(n). — Mateen Ulhaq
– Mateen Ulhaq, Commented Nov 26, 2018 at 6:47

ShlomiF · Accepted Answer · 2018-11-26 06:59:12Z

11

Use numpy.roll. For instance, for the first output you can roll 1 index to the right, meaning along axis 1:

import numpy as np
x = np.array([[0,1,2], [3,4,5], [6,7,8]])
x_shifted = np.roll(x, shift=1, axis=1)

Due to commutativity you can roll twice (once along each dimension) for the two-directional cyclic permutation effect:

x_double_shifted = np.roll(np.roll(x, shift=2, axis=1), shift=2, axis=0)

Obviously can be done more "pretty" ;-)
Good luck!

edited Nov 26, 2018 at 6:59

answered Nov 26, 2018 at 6:43

ShlomiF

2,9451 gold badge16 silver badges21 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

How to shift the entire numpy array, with wrapping

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related