Reshaping numpy array and converting columns to rows

Question

I am sorry for the vagueness of the heading. I have the following task that I can't seem to solve. I have an array of shape (4,2,k). I want to reshape it to shape (2,k*4) while converting columns to rows.

Here is a sample input array where k=5:

sample_array = array([[[2., 2., 2., 2., 2.],
                       [1., 1., 1., 1., 1.]],

                      [[2., 2., 2., 2., 2.],
                       [1., 1., 1., 1., 1.]],

                      [[2., 2., 2., 2., 2.],
                       [1., 1., 1., 1., 1.]],

                      [[2., 2., 2., 2., 2.],
                       [1., 1., 1., 1., 1.]]])

I have managed to get desired output with a for-loop

twos=np.array([])
ones=np.array([])

for i in range(len(sample_array)):
    twos = np.concatenate([twos, sample_array[i][0]])
    ones = np.concatenate([ones, sample_array[i][1]])

desired_array = np.array([twos, ones])

where desired array looks like this:

array([[2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2.,
        2., 2., 2., 2.],
       [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.,
        1., 1., 1., 1.]])

Is there a more elegant solution to my task? I have tried .reshape(), .transpose(), .ravel() but never seem to get the desired outcome.

I am sorry if this is a duplicate, I have looked through a few dozen StackOverflow Qs but found no solution.

@hpaulj thank you. I did, but didn't manage to get desired result. — Julia
– Julia, Commented Sep 4, 2022 at 14:16
The hstack treats the array as 4 arrays of shape (2,5), and concatenates them on the last dimension. The swapaxes/transpose approach makes a (2,4,5) array, and reshapes that to (2,20). Your attempt made 2 arrays of length 20, and joined them. — hpaulj
– hpaulj, Commented Sep 4, 2022 at 15:27

mozway · Accepted Answer · 2022-09-04 14:10:43Z

3

You can swapaxes and reshape:

sample_array.swapaxes(0, 1).reshape(sample_array.shape[1], -1)

output:

array([[2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2.,
        2., 2., 2., 2.],
       [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.,
        1., 1., 1., 1.]])

answered Sep 4, 2022 at 14:10

mozway

267k13 gold badges56 silver badges106 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

ThePyGuy · Accepted Answer · 2022-09-04 14:33:16Z

2

I think simple horizontal stack hstack should work fine:

>>> np.hstack(sample_array)

array([[2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2., 2.,
        2., 2., 2., 2.],
       [1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1., 1.,
        1., 1., 1., 1.]])

answered Sep 4, 2022 at 14:33

ThePyGuy

18.5k5 gold badges24 silver badges55 bronze badges

Collectives™ on Stack Overflow

Reshaping numpy array and converting columns to rows

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related