python - numpy reverse multidimensional array

24
votes

What is the simplest way in numpy to reverse the most inner values of an array like this:

array([[[1, 1, 1, 2],
    [2, 2, 2, 3],
    [3, 3, 3, 4]],

   [[1, 1, 1, 2],
    [2, 2, 2, 3],
    [3, 3, 3, 4]]])

so that I get the following result:

array([[[2, 1, 1, 1],
    [3, 2, 2, 2],
    [4, 3, 3, 3]],

   [[2, 1, 1, 1],
    [3, 2, 2, 2],
    [4, 3, 3, 3]]])

Thank you very much!

2
pythonmultidimensional-arraynumpy

2 Answers

39
votes

How about:

import numpy as np
a = np.array([[[10, 1, 1, 2],
               [2, 2, 2, 3],
               [3, 3, 3, 4]],
              [[1, 1, 1, 2],
               [2, 2, 2, 3],
               [3, 3, 3, 4]]])

and the reverse along the last dimension is:

b = a[:,:,::-1]

or 

b = a[...,::-1]

although I like the later less since the first two dimensions are implicit and it is more difficult to see what is going on.

1
votes

For each of the inner array you can use fliplr. It flips the entries in each row in the left/right direction. Columns are preserved, but appear in a different order than before.

Sample usage:

import numpy as np
initial_array = np.array([[[1, 1, 1, 2],
                          [2, 2, 2, 3],
                          [3, 3, 3, 4]],
                         [[1, 1, 1, 2],
                          [2, 2, 2, 3],
                          [3, 3, 3, 4]]])
index=0
initial_shape = initial_array.shape
reversed=np.empty(shape=initial_shape)
for inner_array in initial_array:
    reversed[index] = np.fliplr(inner_array)
    index += 1

printing reversed

Output:

array([[[2, 1, 1, 1],
        [3, 2, 2, 2],
        [4, 3, 3, 3]],
       [[2, 1, 1, 1],
        [3, 2, 2, 2],
        [4, 3, 3, 3]]])

Make sure your input array for fliplr function must be at least 2-D.

Moreover if you want to flip array in the up/down direction. You can also use flipud