The order of axis when printing a NumPy array

Question

I am working with 3d arrays in NumPy, and I have to look at the array elements a lot. Say I have array a where a.shape is (10,5,3). I am using this is the sense of 3 planes of shape (10,5), and I would like to view it as such.

For example if I

print(a)

I get

[[[ 0  0  0]
  [ 0  0  0]
  [ 0  0  0]
  [ 0  0  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [ 0 22  0]
  [11 22 33]
  [ 0  0  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22 33]
  [ 0  0  0]
  [11 22 33]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22 33]
  [ 0  0  0]
  [11 22  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22 33]
  [11  0  0]
  [11 22  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22 33]
  [ 0 22  0]
  [11  0  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22 33]
  [ 0  0  0]
  [11 22  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22 33]
  [ 0  0  0]
  [11 22 33]
  [ 0  0  0]]

 [[ 0  0  0]
  [11 22  0]
  [ 0 22 33]
  [11 22  0]
  [ 0  0  0]]

 [[ 0  0  0]
  [ 0  0  0]
  [ 0  0  0]
  [ 0  0  0]
  [ 0  0  0]]]

Whereas if I iterate through axis 2 and print like this, it is displayed like I want it:

>>> for i in range(a.shape[2]):
...     print(str(a[:,:,i]) + "\n")
...
[[ 0  0  0  0  0]
 [ 0  0 11  0  0]
 [ 0 11  0 11  0]
 [ 0 11  0 11  0]
 [ 0 11 11 11  0]
 [ 0 11  0 11  0]
 [ 0 11  0 11  0]
 [ 0 11  0 11  0]
 [ 0 11  0 11  0]
 [ 0  0  0  0  0]]

[[ 0  0  0  0  0]
 [ 0 22 22  0  0]
 [ 0 22  0 22  0]
 [ 0 22  0 22  0]
 [ 0 22  0 22  0]
 [ 0 22 22  0  0]
 [ 0 22  0 22  0]
 [ 0 22  0 22  0]
 [ 0 22 22 22  0]
 [ 0  0  0  0  0]]

[[ 0  0  0  0  0]
 [ 0  0 33  0  0]
 [ 0 33  0 33  0]
 [ 0 33  0  0  0]
 [ 0 33  0  0  0]
 [ 0 33  0  0  0]
 [ 0 33  0  0  0]
 [ 0 33  0 33  0]
 [ 0  0 33  0  0]
 [ 0  0  0  0  0]]

Which makes a lot more sense in my mind.

So it seems like my intuition for which way the axes should be visualized is reversed, and my array a should be of the shape (3,5,10). However, I have a whole project going around arrays of a certain format so I can't really change it at this point. I think of the order of the axes as arbitrary, and that's why I feel kind of restricted by this. So is there some quick or even built-in way I can get NumPy to print the axes in reverse, or arbitrary order? I checked out the documentation for numpy.set_printoptions and found nothing of the sort. I realize printing is kind of trivial, but I want to get around all the slicing I have to do to get a good look at my data. Or maybe someone can explain why it's not completely arbitrary and maybe I should think about array axes in a different way.

Answer 1

Unfortunately that is something you can't just do with a simple switch in NumPys print-options. You really have to reorder the axes to get the desired output (which is pretty easy).

If it's just for printing, you can simply define a thin wrapper around the print function:

import numpy as np

def my_numpy_print(arr):
    if arr.ndim == 3:
        print(arr.transpose(2, 0, 1))
    else:  # no idea how you want it to be displayed for different dimensional arrays.
        print(arr)

arr = ...
my_numpy_print(arr)

Output:

[[[ 0  0  0  0  0]
  [ 0  0 11  0  0]
  [ 0 11  0 11  0]
  [ 0 11  0 11  0]
  [ 0 11 11 11  0]
  [ 0 11  0 11  0]
  [ 0 11  0 11  0]
  [ 0 11  0 11  0]
  [ 0 11  0 11  0]
  [ 0  0  0  0  0]]

 [[ 0  0  0  0  0]
  [ 0 22 22  0  0]
  [ 0 22  0 22  0]
  [ 0 22  0 22  0]
  [ 0 22  0 22  0]
  [ 0 22 22  0  0]
  [ 0 22  0 22  0]
  [ 0 22  0 22  0]
  [ 0 22 22 22  0]
  [ 0  0  0  0  0]]

 [[ 0  0  0  0  0]
  [ 0  0 33  0  0]
  [ 0 33  0 33  0]
  [ 0 33  0  0  0]
  [ 0 33  0  0  0]
  [ 0 33  0  0  0]
  [ 0 33  0  0  0]
  [ 0 33  0 33  0]
  [ 0  0 33  0  0]
  [ 0  0  0  0  0]]]