Assigning values to a block in a numpy array

Question

I am trying to change a block within a two-dimensional numpy array by inserting pasting another 2-dim array. The sample below gives me unexpected behavior:

import numpy as np
M=np.ones((4,4))
print(M)
S=[0,1]
print('to be set to zero: ',M[S,:][:,S])
M[S,:][:,S]=np.zeros((2,2))
print('after setting to zero: ',M)

I would expect the upper left corner of M to become zeros. However I get

[[ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]]
to be set to zero:  [[ 1.  1.]
 [ 1.  1.]]
after setting to zero:  [[ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]
 [ 1.  1.  1.  1.]]

It seems that I can extract the upper left block but not write to it. I get the expected behavior if I replace the line

M[S,:][:,S]=np.zeros((2,2))

with

M[0:2,:][:,0:2]=np.zeros((2,2))

What am I doing wrong? Thank you

You're setting entries on a copy.

user2357112
– user2357112

2017-06-04 18:44:53 +00:00
Commented Jun 4, 2017 at 18:44 — user2357112
– user2357112, Commented Jun 4, 2017 at 18:44
M[S,:] creates a copy; M[0:2,:] a view.

hpaulj
– hpaulj

2017-06-04 19:34:25 +00:00
Commented Jun 4, 2017 at 19:34 — hpaulj
– hpaulj, Commented Jun 4, 2017 at 19:34

Allen Qin · Accepted Answer · 2017-06-04 20:01:11Z

3

You can use numpy advanced indexing ix_

M[np.ix_(S,S)]=0

M
Out[622]: 
array([[ 0.,  0.,  1.,  1.],
       [ 0.,  0.,  1.,  1.],
       [ 1.,  1.,  1.,  1.],
       [ 1.,  1.,  1.,  1.]])

answered Jun 4, 2017 at 20:01

Allen Qin

20k9 gold badges55 silver badges68 bronze badges

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

Collectives™ on Stack Overflow

Assigning values to a block in a 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