Find N highest values per row of a 2d numpy array

Question

I have a 2d numpy array a:

a = np.array(range(0,25)).reshape(5,5)
---
[[ 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]]

I want to find maxmium N values of each row and replace them with 100. I do this in a slow way:

N = 2
idx = a.argsort()
for i in range(a.shape[0]):
    a[i,idx[i][::-1][0:N]] = 100
print(a)
---
[[  0   1   2 100 100]
 [  5   6   7 100 100]
 [ 10  11  12 100 100]
 [ 15  16  17 100 100]
 [ 20  21  22 100 100]]

Actually the shape of my matrix is 6000*6000. How to do this in a better way? Like apply?

yatu · Accepted Answer · 2020-05-18 08:47:59Z

1

You can use argpartition here:

N=2
ix = a.argpartition(-N)[:,-N:]
a[np.arange(a.shape[0])[:,None], ix] = 100

print(a)
array([[  0,   1,   2, 100, 100],
       [  5,   6,   7, 100, 100],
       [ 10,  11,  12, 100, 100],
       [ 15,  16,  17, 100, 100],
       [ 20,  21,  22, 100, 100]])

edited May 18, 2020 at 8:47

answered May 18, 2020 at 8:42

yatu

88.7k12 gold badges93 silver badges148 bronze badges

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

Collectives™ on Stack Overflow

Find N highest values per row of a 2d numpy array

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related