python populate elements in array

Question

I am very new to Python (previous Matlab user).

I have the array

y_pred =  [None] * 128

test_idx is an array of indeces

array([  3,   4,   5,  19,  28,  30,  38,  39,  47,  49,  50,  51,  54,
        64,  74,  81,  84,  85,  90,  91,  93,  97, 102, 103, 106, 107,
       109, 111, 115, 121], dtype=int64)

I would like to replace the values of y_pred corresponding to the test_idx with the array results

array([0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0,
       0, 0, 0, 0, 0, 0, 0])

if I try

y_pred[test_idx] = results

I get the error: TypeError: only integer arrays with one element can be converted to an index

y_pred is not an array, but a list, which can't be indexed like that. Perhaps use an array instead? — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Nov 10, 2016 at 9:27
Also, you should really use the numpy tag is you are going to ask a question about numpy arrays. — juanpa.arrivillaga
– juanpa.arrivillaga, Commented Nov 10, 2016 at 9:37
Do you really want y_pred`` to be a list of None`s? Not array of zeros? What's the Matlab equivalent? — hpaulj
– hpaulj, Commented Nov 10, 2016 at 13:36
Show exactly what you expect. There's too much ambiguity in the current example. — hpaulj
– hpaulj, Commented Nov 10, 2016 at 13:59

pt12lol · Accepted Answer · 2016-11-10 09:27:45Z

1

Your y_pred is a native Python list and you cannot index it by anything other than integers. You should create it as a numpy array.

Try to initialize it the way below:

import numpy as np
y_pred = np.array([None] * 128)

answered Nov 10, 2016 at 9:27

pt12lol

2,4411 gold badge26 silver badges51 bronze badges

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

7 Comments

juanpa.arrivillaga Over a year ago

You have to consider carefully whether or not to use None, which will force y_pred to have object dtype, which kills performance.

juanpa.arrivillaga Over a year ago

True, but it is a consideration worth mentioning, especially if the OP is new to python/numpy.

pt12lol Over a year ago

yup, thanks for your comments, this is important thing to mention

woockashek Over a year ago

@pt12lol: I believe that you are adressing completely different issue. Indexing with something different than integer is not a problem here.

juanpa.arrivillaga Over a year ago

@woockashek No, the question is clearly related to numpy.

|

Moamen · Accepted Answer · 2016-11-10 09:32:34Z

1

I see you are doing a lot of matlab here ;)

If you want to do it the pythonic way! you need to loop. Like:

for index, value in zip(test_idx, results):
    y_pred[index] = value

answered Nov 10, 2016 at 9:32

Moamen

6987 silver badges20 bronze badges

Comments

TheBlackCat · Accepted Answer · 2016-11-13 19:29:37Z

Short version: replace y_pred = [None] * 128 with y_pred = np.full(128, np.NaN). Everything else will work once you do that.

Long version:

The problem as others have said, is that y_pred is a list, not an array. Lists do not support getting or setting multiple indexes at a time, you need to use a numpy array for that.

The simplest approach to what you want is to use a numpy array of what is called a "sentinel" value, a value that indicates that nothing was written there. The most obvious choice for this value is probably either 0 or NaN, but depending on what you want to do with the results it could be just about anything. NaN is the closets to None you will get in a numerical numpy array, but 0 can make some later calculations easier (for example NaN != NaN, you need to use np.isnan to identify NaN values later).

To do this with zeros, you can replace y_pred = [None] * 128 with y_pred = np.zeros(128). To use NaN (or any other value), you can use y_pred = np.full(128, np.NaN) (there is also a np.ones but I don't think it will help you here). The rest of your code can then be used as-is.

woockashek · Accepted Answer · 2016-11-10 09:32:43Z

0

you could do it like this:

for i in xrange(len(test_idx)):
    y_pred[test_idx[i]] = results[i]

or:

for key, val in zip(test_idx, results):
    y_pred[key] = val

I assume your both lists have the same length.

answered Nov 10, 2016 at 9:32

woockashek

1,67412 silver badges27 bronze badges

Collectives™ on Stack Overflow

python populate elements in array

4 Answers 4

7 Comments

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

7 Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related