Object initialization in Python

Question

I've created a simple class with object that has init value x0. When I change another value x, my x0 is also changing.

I thought that x0 should remain unchanged. Could you please explain why is this happening?

file main.py:

import numpy as np
from simpleclass import test


def main():

    params = dict()
    params['position'] = np.array([1.0, 2.0])

    object = test(params)

    print(object.x0)
    print(object.x)

    object.run(2)

    print(object.x0)
    print(object.x)


if __name__ == "__main__":
    main()

file simpleclass.py:

class test():

    def __init__(self, params):
        self.x0 = params['position']
        self.x = self.x0

    def run(self, num):
        self.x += self.x*num

result:

[ 1.  2.]
[ 1.  2.]
[ 3.  6.]
[ 3.  6.]

score 3 · Accepted Answer · answered Oct 06 '17 at 11:58

problem is with

class test():
    def __init__(self, params):
        self.x0 = params['position']
        self.x = self.x0

    def run(self, num):
        self.x += self.x*num

self.x = self.x0 here self.x and self.x0 are pointing same object. you can make copy of self.x0.

import copy

class test():
    def __init__(self, params):
        self.x0 = params['position']
        self.x = copy.deepcopy(self.x0)

    def run(self, num):
        self.x += self.x*num

score 1 · Answer 2 · answered Oct 06 '17 at 12:04

1

surya singh is right, just print the memory address, you will get the same numbers

class test():
    def __init__(self, params):
        self.x0 = params['position']
        self.x = self.x0
        print id(self.x)
        print id(self.x0)

answered Oct 06 '17 at 12:04

Jiri Semmler

411
3
11

Thanks! I will keep this in mind – Gatto Nou Oct 06 '17 at 12:11

score 1 · Answer 3 · answered Oct 06 '17 at 12:38

You are only passing a reference

By self.x = self.x0, you are only passing a reference. With self.x += self.x*num, the reference is unchanged too. So after those two operations, x and x0 still points to the same array.

It would be different if you used an immutable object, such as tuple

params['position'] = (1, 2)

With tuple, += does something different than what you want, but you see, x and x0 points to different object.

(1, 2)
(1, 2)
(1, 2)
(1, 2, 1, 2, 1, 2)

Use a copy

You want to create a copy of your array, numpy has a built-in method for that

self.x = self.x0.copy()

Result:

[ 1.  2.]
[ 1.  2.]
[ 1.  2.]
[ 3.  6.]

Object initialization in Python

3 Answers3

You are only passing a reference

Use a copy