Randomly replace cells in a numpy array without replacement

Question

I have a 25 x 25 numpy array filled with 0s. I want to randomly replace 0s within the array with 1s, with the number of 0s to be replaced defined by the user and ranging from 1 to 625. How can I do this without replacement becoming an issue given there are two coordinate values, x and y?

Code for generating a 25 x 25 array:

 import numpy as np
 Array = np.zeros((25,25))

Expected result on a 4 x 4 array:

Before

0,0,0,0
0,0,0,0
0,0,0,0
0,0,0,0

After (replace 1)

0,0,0,0
0,0,1,0
0,0,0,0
0,0,0,0

After (replace 16)

1,1,1,1
1,1,1,1
1,1,1,1
1,1,1,1

Asked By: JoshuaAJones

||

Source

Answer 1

Try:

import random

import numpy as np

SIZE = 25

howmany = 500

array = np.zeros((SIZE, SIZE), dtype=int)
xyset = set()

for i in range(SIZE):
    for j in range(SIZE):
        xyset.add((i, j))

for num in range(howmany):
    xy = random.choice(list(xyset))
    array[xy[0], xy[1]] = 1

    xyset.discard(xy)

print(array)

Answered By: user19077881

Answer 2

Notes:

One may generate a unique sequence of random integers from 0 to 624, and mutate the corresponding element of the flattened array according to the generated sequence. Working with the flattened array, in this specific case, enables the development of a (relatively) concise code.
One can use np.random.choice together with the argument replace=False to generate a non-repetitive sequence of random numbers. You can see more information on NumPy docs.

Here is one of the solutions for a case with a 2D array of the size of 5×5 that undergoes 5 mutations:

import numpy as np

dim = 5     #<========== size of each axis of array -- in your case, it would be 25
num_mut = 5 #<========== number of mutations -- in your case it would be an integer from 0 to 624

Array = np.zeros((dim,dim))
Array_flat = Array.flatten()
print('before mutation:')
print(Array)

ind_mut = np.random.choice(dim*dim, size=num_mut, replace=False).tolist()
Array_flat[ind_mut]=1
Array_mut = Array_flat.reshape(Array.shape)

print('nafter mutation:')
print(Array_mut)

Output:

Answered By: learner

Randomly replace cells in a numpy array without replacement

Question:

Answers: