Numpy reorganize array

Question

Let's say I have got the following numpy array

A = 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],[25,26,27,28,29]])

out[]: 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, 27, 28, 29]])

I would like to reorganize it in such a way that the output is

out[] : array([[0,5],
         [1,6],
         [2,7],
         [3,8],
         [4,9],
         [10,15],
         [11,16],
         [12,17],
         [13,18],
         [14,19],
         [20,25],
         [21,26],
         ....,
         [24,29]])

I have been trying different combinations of np.reshape, tranpose, flatten, np.swapaxes, but with no success.

The real array has tens or sometimes hundreds of rows.

Originally, the data is given as DataFrame, but I realized that converting to numpy array could be a better alternative... Can it be done directly using pandas?

So the first, third, etc. row should be the first column and the second, fourth, etc. row should be the second column? — bayes2021, Jan 11 '23 at 06:45
Yes, in general for a, let's say 100x5 dimensional array, I would like to group every 10 rows and put them as columns (this subgroup would be 5x10), then the 10 subgroups are later appended, such that I end up with a 50x10 dimensional array — ktitimbo, Jan 11 '23 at 06:58
@KelvinTitimbo You can refer this answer for general logic to solve these kind of problems https://stackoverflow.com/a/47978032/5462372 — MSS, Jan 11 '23 at 09:43

MSS · Answer 1 · 2023-01-12T06:42:52.203

Yes this can be done in this way using as_strided.

from numpy.lib.stride_tricks import as_strided

A = 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],[25,26,27,28,29]])
A = A.reshape(3, 2, 5)
#print(A)
S = A.itemsize
out = as_strided(A, shape=(3,5,2), strides=(2*A.shape[1]*S ,S, A.shape[1]*S)).copy()
out = out.reshape(15,2)
print(out)
>> [[ 0  5]
 [ 1  6]
 [ 2  7]
 [ 3  8]
 [ 4  9]
 [10 15]
 [11 16]
 [12 17]
 [13 18]
 [14 19]
 [20 25]
 [21 26]
 [22 27]
 [23 28]
 [24 29]]

Numpy reorganize array

1 Answers1