Modify actual element value in array numpy

Question

I can't modify the actual value of a numpy array in a loop. My code is the following :

labels_class = np.copy(labels_train)
for label in labels_class:
  labels_class[label] = 1 if (label == classifier) else 0

labels_class - is just a numpy array of size N and of values [0, 39]. The value of labels_class[label] is correct(==modified) in the loop, but outside of the loop labels_classremains unchanged.

I have also tried nditer, did not work :

 for label in np.nditer(labels_class, op_flags=['readwrite']):
      label = 1 if (label == classifier) else 0

In the reference, it is said that "to actually modify the element of the array, x should be indexed with the ellipsis"

How do I do that? What is the syntax?

Perhaps you meant to enumerate through the numpy array. Simply iterating over an iterable yields the elements of the iterable, not the indexes of the elements. — Shashank
– Shashank, Commented Apr 6, 2015 at 20:05
I meant accessing the elements of the array in the loop and modifying actual values of it. — Uylenburgh
– Uylenburgh, Commented Apr 6, 2015 at 20:09
To actually change label you have to use something like label[:]=... or label[...] = .... Review the nditer tutorial if you want to go that route. — hpaulj
– hpaulj, Commented Apr 6, 2015 at 21:45
docs.scipy.org/doc/numpy/reference/… 'modifying array values with nditer' — hpaulj
– hpaulj, Commented Apr 6, 2015 at 21:51

sedavidw · Accepted Answer · 2015-04-06 20:21:53Z

7

Your iterator is not creating indices, but the actual elements in the array

for label in labels_class

In the above label is not an index, but the actual element you are trying to change

You can do something like this:

for i, label in enumerate(labels_class):
     labels_class[i] = 1 if (label == classifier) else 0

answered Apr 6, 2015 at 20:21

sedavidw

11.8k18 gold badges70 silver badges103 bronze badges

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

3 Comments

Uylenburgh Over a year ago

This is the solution! Thank you. However, I must say that in for label in labels_class label is not the actual element but a copy.

Bi Rico Over a year ago

You don't need the 1 if ... else 0 part, you can just leave label == classifier.

Bi Rico Over a year ago

@Oleksandra, python variables are different than c variables, for example. Python ints and floats are immutable so there is no difference between a copy and the actual element.

hpaulj · Accepted Answer · 2015-04-06 21:56:07Z

5

The syntax for modifying array values with nditer is demonstrated at

http://docs.scipy.org/doc/numpy/reference/arrays.nditer.html#modifying-array-values

a = np.arange(6).reshape(2,3)
for x in np.nditer(a, op_flags=['readwrite']):
    x[...] = 2 * x

'x should be indexed with the ellipsis' refers to the x[...].

Indexing with enumerate is perfectly fine as well. But this is the way to do it with nditer. See later sections in the nditer page about using flags=['f_index'].

When iterating over arrays you need to clearly understand the difference between variables which are indexes, scalars, or array elements which can be modified. x = 1 is not the same as A[i]= 1 or x[...]=1.

answered Apr 6, 2015 at 21:56

hpaulj

233k14 gold badges260 silver badges392 bronze badges

Comments

jliu1999 · Accepted Answer · 2020-12-18 19:11:28Z

0

how about this?

labels_class[label_class == classifier] = 1
labels_class[label_class != classifier] = 0

answered Dec 18, 2020 at 19:11

jliu1999

124 bronze badges

Collectives™ on Stack Overflow

Modify actual element value in array numpy

3 Answers 3

3 Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

3 Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related