Mutable and Immutable Objects

An object is mutable when its contents can change after creation, and immutable when they cannot. Among the built-in types, lists, dictionaries, and sets are mutable; numbers, strings, tuples, and frozensets are immutable. Operations do not modify the immutable object itself, but need not allocate a new object. Use the returned result of text.replace(), for example by assigning it, while items.append() changes the existing list.

Assignment binds a name to an object and never copies it. After alias = original, both names refer to one object, so a mutation through either is visible through both, and original is alias is True. Strings cannot change in place, but an immutable container can refer to mutable objects: a list inside a tuple can still change. Tuple conversion is not a deep freeze. Rebinding one name, alias = [9], only changes where that name points.

To work on an independent collection, copy it: list(original), original.copy(), and original[:] all produce a new list. These are shallow copies, so the new container refers to the same inner objects and mutating a nested list is still visible from both. copy.deepcopy() recursively copies supported nested objects, at a time and memory cost; it preserves shared relationships through its memo and is not a universal clone of files or external resources.

Two consequences appear often in real code. First, only hashable objects can be dictionary keys or set elements, and built-in lists, dictionaries, and sets are unhashable. Hashability requires a stable hash consistent with equality; it is not synonymous with immutability. A tuple containing a list is immutable but unhashable, while custom objects may be mutable and hashable. Second, a default argument is evaluated once when the function is defined, so a mutable default such as def add(item, target=[]) accumulates values across calls; use None as the default and create the list inside the function. When passing a collection to code you do not control, consider passing a copy or an immutable type.

References: Objects, values and types, copy module. See it in use in Python Collections and Comprehensions.