Rotate Matrix using Python

Question

This is my solution to Exercise 1.7 from Cracking the Coding Interview. I would appreciate any feedback on coding style and algorithm efficiency.

I do know that there is an existing function in numpy for rotating a matrix, however I am trying to implement this as an exercise.

The problem statement follows:

Given an image represented by an NxN matrix where each pixel in the image is 4 bytes, write a method to rotate the image by 90 degrees. Can you do this in place?

import unittest
def rotate_square_matrix_right_90(matrix: list) -> list:
 """Rotate an NxN matrix 90 degrees clockwise."""
 n = len(matrix)
 for layer in range((n + 1) // 2):
 for index in range(layer, n - 1 - layer, 1):
 matrix[layer][index], matrix[n - 1 - index][layer], \
 matrix[index][n - 1 - layer], matrix[n - 1 - layer][n - 1 - index] = \
 matrix[n - 1 - index][layer], matrix[n - 1 - layer][n - 1 - index], \
 matrix[layer][index], matrix[index][n - 1 - layer]
 return matrix
class MyTest(unittest.TestCase):
 def test_rotate_matrix_right_90(self):
 input_matrix = [[0, 2, 4, 6], [-1, 1, 3, 5], [-2, 0, 2, 4], [-3, -1, 1, 3]]
 correct_rotated_matrix = [[-3, -2, -1, 0], [-1, 0, 1, 2], [1, 2, 3, 4], [3, 4, 5, 6]]
 self.assertSequenceEqual(rotate_square_matrix_right_90(input_matrix), correct_rotated_matrix)
if __name__ == "__main__":
 unittest.main()

Answer 1

You can use much simpler algorithm in python: Transpose matrix:

zip(*matrix)

Inverse rows in transposed matrix (equals rotation right):

list(list(x)[::-1] for x in zip(*matrix))

However, if you want to rotate left, you need first inverse rows and then transpose, which is slightly more code.

• jayko03: To rotate left side, list(list(x) for x in zip(*matrix))[::-1]
• Paul Lan: Try it online from my sample code: Rotate left, rotate right: trinket.io/python/f7ad7f9864