References, (Im)mutability

Immutability: Numbers

Numbers are immutable …

• Object of type int with value 42

• Variable a points to it (“gives it a name”)

• The object cannot change its value - there is no method to modify an integer object

• ⟶ The latter situation is equivalent to the former (which is the implementation)

a = 42 b = a

a = 42 b = 42

Immutability: Tuples

Same with tuples

Like lists, but immutable No way to modify a tuple No appending No slice assignment No nothing So both of these are equivalent To the user, b is a copy of a

>>> a = (42, "xy", 13) >>> b = a

Mutability: Lists (1)

Lists are mutable …

>>> a = [1, 2, 3] >>> b = a >>> b [1, 2, 3] >>> b.append(4) >>> b [1, 2, 3, 4] >>> a [1, 2, 3, 4]

Objects can be modified E.g. by using append()

Mutability: Lists (2)

Danger …

• Take care when passing complex data structures

• Not passed by copy (as in C++)

• Passed by reference (as in Java)

• Make a copy if needed

Copying a list

>>> a = [1, 2, 3]
>>> b = a[:]
>>> a.append(4)
>>> b
[1, 2, 3]

Shallow Copy

>>> a = [1, [1, 2, 3], 2] >>> b = a[:] >>> b [1, [1, 2, 3], 2] >>> a[1].append(4) >>> a [1, [1, 2, 3, 4], 2] >>> b [1, [1, 2, 3, 4], 2] >>> a[1] is b[1] True

Only first level copied Shallow copy a[1] is a reference is: object identity

Deep Copy

Solution: not easy

• Recursive structure traversal

• Handling every possible type

• Dedicated module in the standard library: copy

>>> import copy
>>> a = [1, [1, 2, 3], 2]
>>> b = copy.deepcopy(a)
>>> a[1] is b[1]
False