numpy is double transposition necessary in this specific case?

I have an array

xx = np.arange(24).reshape(2, 12)

array([[ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11],
       [12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]])

and I would like to reshape it, to obtain

array([[[ 0,  1,  2,  3],
        [12, 13, 14, 15]],

       [[ 4,  5,  6,  7],
        [16, 17, 18, 19]],

       [[ 8,  9, 10, 11],
        [20, 21, 22, 23]]])

I can achieve it via

xx.T.reshape(3, 4, 2).transpose(0, 2, 1)

But it has to be transposed twice, which seems unnecessary to me. So could somebody confirm that this is the only way of doing it or provide more readable solution otherwise?
Thanks!

>Solution :

It is possible to do a single transpose:

data = np.arange(24).reshape(2, 12)
data = data.reshape(2, 3, 4).transpose(1, 0, 2)

Edit:

I checked this using itertools.permutations and itertools.product:

import itertools
import numpy as np

data = np.arange(24).reshape(2, 12)
desired_data = np.array([[[ 0,  1,  2,  3],
                          [12, 13, 14, 15]],
                         
                         [[ 4,  5,  6,  7],
                          [16, 17, 18, 19]],
                         
                         [[ 8,  9, 10, 11],
                          [20, 21, 22, 23]]])

shapes = [2, 3, 4]
transpose_dims = [0, 1, 2]

shape_permutations = itertools.permutations(shapes)
transpose_permutations = itertools.permutations(transpose_dims)

for shape, transpose in itertools.product(
    list(shape_permutations),
    list(transpose_permutations),
):
    
    new_data = data.reshape(*shape).transpose(*transpose)

    try:
        np.allclose(new_data, desired_data)
    except ValueError as e:
        pass
    else:
        break

print(f"{shape=}, {transpose=}")

shape=(2, 3, 4), transpose=(1, 0, 2)