How to flip numpy array along the diagonal efficiently?

Question

Lets say that i have the following array (note that there is a 1 in the [2,0] position and a 2 in the [3,4] position):

[0, 0, 0, 0, 0]
[0, 0, 0, 0, 0]
[1, 0, 0, 0, 0]
[0, 0, 0, 0, 2]
[0, 0, 0, 0, 0]

and I want to flip it along the diagonal efficiently such that:

[0, 0, 1, 0, 0]
[0, 0, 0, 0, 0]
[0, 0, 0, 0, 0]
[0, 0, 0, 0, 0]
[0, 0, 0, 2, 0]

This does not work with fliplr or rot90 or flipud. Would like efficient answer rather than just an answer since unfortunately this is not being performed on matrices this small.

https://docs.scipy.org/doc/numpy/reference/generated/numpy.ndarray.transpose.html — Arne, Apr 14 '20 at 00:00
Does this answer your question? [Matrix Transpose in Python](https://stackoverflow.com/questions/4937491/matrix-transpose-in-python) — AMC, Apr 14 '20 at 02:07

score 16 · Accepted Answer · edited Mar 26 '21 at 09:43

16

Both np.rot90(np.fliplr(x)) and transposing the array solves this.

a = np.random.uniform(size=(5,5))
a.T == np.rot90(np.fliplr(a))

edited Mar 26 '21 at 09:43

Jurgen Strydom

3,540
1
23
30

answered Apr 14 '20 at 00:04

negfrequency

1,801
3
18
30

How to flip numpy array along the diagonal efficiently?

1 Answers1

Linked