CODEWITHSUNDEEP
pythonarraysnumpynumpy-ndarraynumpy-slicing
Enger Bewza
Enger Bewza

Reputation: 103

Angled slice of 3D array

In NumPy I understand how to slice 2D arrays from a 3D array using this article:

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]
          [24 25 26]]]

Slicing would give me:

i_slice = array[0]

    [[0  1  2]
     [3  4  5]
     [6  7  8]]

j_slice = array[:, 0]

    [[0  1  2]
     [9  10 11]
     [18 19 20]]

k_slice = array[:, :, 0]

    [[0  3  6]
     [9  12 15]
     [18 21 24]]

But is it possible to slice at a 45 degree angle? Such as:

j_slice_down = array[slice going down starting from index 0]

    [[0  1  2]
     [12 13 14]
     [24 25 26]]

I can do this on all 3 axis' going up or down and wrapping around with lists and for loops, but there must be a better way in NumPy.

Upvotes: 8

Views: 2910

Answers (2)

hpaulj
hpaulj

Reputation: 231335

In [145]: arr[np.arange(3), np.arange(3),:]
Out[145]: 
array([[ 0,  1,  2],
       [12, 13, 14],
       [24, 25, 26]])

Upvotes: 3

Chris
Chris

Reputation: 29732

You can try with np.diagonal:

arr = np.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],
          [24 ,25 ,26]]])

np.diagonal(arr).T
array([[ 0,  1,  2],
       [12, 13, 14],
       [24, 25, 26]])

Upvotes: 2

Related Questions