NP reshape (add extra 2 dimension)

Question

I will illustrate my question with an example. If I have the array:

a = np.array([[1,2,3,4],
              [9,10,11,12],
              [17,18,19,20],
              [25,26,27,28]])

I would like to get

array([[[ 1,  2],   [[ 3,  4],
        [9, 10],    [11, 12],
        [17, 18],    [19, 20],
        [25, 26]]    [27, 28]],

So apparently if my array was MxN , now it will be Mx(N/2)x2. How to do it? I tried:

import numpy as np

# pre-computed data
data.reshape(data.shape[0], data.shape[1]//2, 2)

, does not work as expected

Asked By: Petar Ulev

||

Answer 1

Use np.split (on column axis) + np.stack combination:

arr = np.stack(np.split(a, 2, axis=1))
print(arr)

[[[ 1  2]
  [ 9 10]
  [17 18]
  [25 26]]

 [[ 3  4]
  [11 12]
  [19 20]
  [27 28]]]

Answer 2

Alternative solution:

out = a.reshape(4, 2, 2).swapaxes(0, 1)

out:

array([[[ 1,  2],
        [ 9, 10],
        [17, 18],
        [25, 26]],

       [[ 3,  4],
        [11, 12],
        [19, 20],
        [27, 28]]])

Question: