numpy random shuffle by row independently

Question

I have the following array:

 a= array([[  1,  2, 3],
           [  1,  2, 3],
           [  1,  2, 3])

I understand that np.random,sh

Answer 1

You can do it with numpy without any loop or extra function, and much more faster. E. g., we have an array of size (2, 6) and we want a sub array (2,2) with independent random index for each column.

import numpy as np

test = np.array([[1, 1],
                 [2, 2],
                 [0.5, 0.5],
                 [0.3, 0.3],
                 [4, 4],
                 [7, 7]])

id_rnd = np.random.randint(6, size=(2, 2))  # select random numbers, use choice and range if don want replacement.
new = np.take_along_axis(test, id_rnd, axis=0)

Out: 
array([[2. , 2. ],
       [0.5, 2. ]])

It works for any number of dimensions.