Create random bool array

Question

I want to create a randomized bool array with a given length a given number of True values efficiently in Python. I could not find a single command to do so, the following does what I want (twice). Is there anything more elegant way to do it?

import numpy as np

def randbool(length,numtrue):
    index_array=np.random.choice(length,numtrue,replace=False)
    bool_array=np.zeros(length,dtype=bool)
    bool_array[index_array]=True
    return(bool_array)

def randbool2(length,numtrue):
    bool_array=np.hstack((np.tile(True,numtrue),np.tile(False,length-numtrue)))
    np.random.shuffle(bool_array)
    return(bool_array)

print(randbool(5,2))
print(randbool2(5,2))

And no, I don't think that there is a more elegant approach here — Ben Grossmann
– Ben Grossmann, Commented Nov 7, 2022 at 14:37

obchardon · Accepted Answer · 2022-11-07 15:38:19Z

4

Still another option:

def randbool(l,n):
   return np.random.permutation(l)<n

np.random.permutation randomly permute np.arange(l).

edited Nov 7, 2022 at 15:38

answered Nov 7, 2022 at 15:25

obchardon

10.8k1 gold badge19 silver badges36 bronze badges

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

1 Comment

Okapi575 Over a year ago

Cool idea, to compare the permutations with the number of True values

mozway · Accepted Answer · 2022-11-07 14:39:44Z

1

Another suggestion:

def randbool3(length, numtrue):
    return np.random.choice(length, length, replace=False) < numtrue

Similar in efficiency to the first approach but a bit shorter in code

answered Nov 7, 2022 at 14:39

mozway

267k13 gold badges56 silver badges106 bronze badges

Comments

Sam Mason · Accepted Answer · 2022-11-07 15:54:15Z

1

Both of your examples have small n and is therefore in the regime where native Python code can be faster than Numpy. For example, given:

from random import shuffle

def randbool4(length, numtrue):
    arr = [False] * length
    arr[:numtrue] = [True] * numtrue
    shuffle(arr)
    return arr

I can execute randbool4(5,2) in ~4µs. Your variants take ~30µs and mozway's are similar, while obchardon's Numpy variant takes only ~6µs which is surprisingly close.

Obviously when n gets large then Numpy's native kernels of win, but thought it worth pointing out.

answered Nov 7, 2022 at 15:54

Sam Mason

16.5k1 gold badge49 silver badges71 bronze badges

Collectives™ on Stack Overflow

Create random bool array

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

1 Comment

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related