How to use a numpy boolean array to modify another numpy array?

Question

I have a boolean array that looks like this:

arr_a = np.array(
[[False, False, False],
[True, True, True],
[True, True, True],
[False, False, False]]
)

and another array that looks like this:

arr_b = np.array(
[[100, 100, 100],
[200, 200, 200]]
)

I am looking for a function that I can call like this: np.boolean_combine(arr_a, arr_b), to return an array that will replace the 1's in arr_a with the values from arr_b, for an end result that looks like this:

np.array(
[[0, 0, 0]
[100, 100, 100],
[200, 200, 200],
[0, 0, 0]]
)

Is there such a function?

Will the shapes always be a match? So will arra_b.size always equal the number of True in arr_a? — user3483203
– user3483203, Commented Jun 26, 2019 at 16:13

yatu · Accepted Answer · 2019-06-26 16:18:48Z

2

You can create a new array of the same dtype as arra_b, take a slice view using arr_a an assign the values from arra_b:

out = arr_a.astype(arra_b.dtype)
out[arr_a] = arra_b.ravel()

array([[  0,   0,   0],
       [100, 100, 100],
       [200, 200, 200],
       [  0,   0,   0]])

answered Jun 26, 2019 at 16:18

yatu

88.6k12 gold badges93 silver badges148 bronze badges

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

Comments

Kamuish · Accepted Answer · 2019-06-26 16:44:32Z

1

If your arr_a is made of 1's and 0's:

import numpy as np 
arr_a = np.array(
[[0, 0, 0],
[1, 1, 1],
[1, 1, 1],
[0, 0, 0]])
arra_b = np.array(

[[100, 100, 100],

[200, 200, 200]])


arr_a[np.where(arr_a)]  = arra_b.reshape(arr_a[np.where(arr_a)].shape)

This works, assuming that the shapes are a match

edited Jun 26, 2019 at 16:44

answered Jun 26, 2019 at 16:19

Kamuish

1186 bronze badges

2 Comments

Preethi Vaidyanathan Over a year ago

Here is what this returned: >>> arr_a = np.array([[0, 0, 0], [1, 1, 1], [0, 0, 0]]) >>> arra_b = np.array([[100, 100, 100],[200, 200, 200]]) >>> arr_a[np.where(arr_a)] = arra_b.reshape(arr_a[np.where(arr_a)].shape) Traceback (most recent call last): File "<stdin>", line 1, in <module> ValueError: cannot reshape array of size 6 into shape (3,)

Kamuish Over a year ago

I just ran it again and it works without an error. Check the edit to my answer to see if my arrays a and b are equal to yours

mao95 · Accepted Answer · 2019-06-26 16:25:34Z

0

You can use zip(). Supposing that arr_a and arr_b are (obviously, from your problem) of the same dimension:

def boolean_combine(arr_a, arr_b) -> np.array:
  combined = []
  for row_a, row_b in zip(arr_a, arr_b):
    row_combined = []
    for a, b in zip(row_a, row_b):
      if a == 'True':
        row_combined.append(b)
      else:
        row_combined.append(a)
    combined.append(np.asarray(row_combined))
  return np.asarray(combined)

Then, you can call this function in your main just by typing combined = boolean_combine(arr_a, arr_b)

edited Jun 26, 2019 at 16:25

answered Jun 26, 2019 at 16:20

mao95

1,12212 silver badges23 bronze badges

Collectives™ on Stack Overflow

How to use a numpy boolean array to modify another numpy array?

3 Answers 3

Comments

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related