Mutable objects in python

Este artículo también está disponible en: es

Let’s assume we have defined the following function:

def f(x, l=[]):
    for i in range(x):
        l.append(i)
    print(l)

If we have to guess what is the output of these calls to the function. A lot of us will think is:

>>> f(2)
[0, 1]
>>> f(3, [3, 4, 5])
[3, 4, 5, 0, 1, 2]
>>> f(3) # WRONG RESULT
[0, 1, 2]

However, the correct result is:

>>> f(2)
[0, 1]
>>> f(3, [3, 4, 5])
[3, 4, 5, 0, 1, 2]
>>> f(3) # CORRECT RESULT
[0, 1, 0, 1, 2]

This is due to the fact that lists in python are mutable objects, and therefore whenever they are defined as default parameters in a function, the list behaves as a global variable common to all function’s executions.

When we call to f(2) for the first time, the value [] is being used by default. But when f(3) is executed the same global variable used when f(2) was called is used, which last value is [0, 1].

Table of objects in python

Mutable objects in python are passed by reference. Immutable objects is like they were being passed by value. An example of that is the following code:

def f(l):
    l.append(23)

>>> l = [37]
>>> f(l)
>>> print(l)
[37, 23]

Share on: Twitter ❄ Facebook ❄ Email

Asier Cardoso Sánchez Matemático e Ingeniero Informático