How to avoid static lists in python classes?

Question

Python

#!/usr/bin/python3
import random

class Vector:
  
  element = []
  N = 0
  
  def __init__(self, N):
    for i in range(N):
      self.element.append(0)
    self.N = N

  def __str__(self):
    out_str = ""
    for i in range(self.N):
      out_str = out_str +  str(self.element[i]) + " "
    return out_str

  def generate(self):
    random.seed()
    for i in range(self.N):
      self.element[i] = float(random.randint(1, 100))

first = Vector(3)
first.generate()
print(first)
second = Vector(3)
second.generate()

print(first)
print(second)

On the output to the console I get:
38.0 41.0 82.0
34.0 42.0 34.0
34.0 42.0 34.0
Why do the values ​​of the list belonging to the first object change when calling the method of the second object? How to avoid?

Answer 1

When you declare member variables, as you did, they are associated with the class, not the object. Similar to C++, you've made the elements and N member variables static.
You need to create member variables in the constructor, not in the class definition:

class Vector:

  def __init__(self, N):
    self.elenent = []
    self.N = N
    for i in range(N):
      self.element.append(0)

  def __str__(self):
    out_str = ""
    for i in range(self.N):
      out_str = out_str +  str(self.element[i]) + " "
    return out_str

  def generate(self):
    random.seed()
    for i in range(self.N):
      self.element[i] = float(random.randint(1, 100))

Or use the __slots__ special variable:

class Vector:

  __slots__ = ['element', 'N']

  def __init__(self, N):
    self.elenent = []
    self.N = N
    for i in range(N):
      self.element.append(0)

  def __str__(self):
    out_str = ""
    for i in range(self.N):
      out_str = out_str +  str(self.element[i]) + " "
    return out_str

  def generate(self):
    random.seed()
    for i in range(self.N):
      self.element[i] = float(random.randint(1, 100))

The difference is that in the second case, you can no longer add other member variables, and the objects will take up less memory.