How to exclude specific references from deepcopy?

Question

I have object which has it's own content (i.e. list of something) and a reference to another object, with which it is linked. How can I exclude the reference to the other object from being deep-copied?

from copy import deepcopy
class Foo:
    def __init__(self, content, linked_to):
        self.content = content
        self.linked_to = linked_to

a1 = Foo([[1,2],[3,4]], None)
a2 = Foo([[5,6],[7,8]], a1)

a3 = deepcopy(a2) # <- I don't want there, that a3.linked_to will be copied
# I want that a3.linked_to will still point to a1

a3.linked_to.content.append([9,10])
print a1.content # [[1,2],[3,4]], but I want [[1,2],[3,4], [9,10]]

That's not a weak reference. That's just a normal reference to another object. — Martijn Pieters, May 15 '18 at 14:05
You can implement it yourself by implementing `__deepcopy__` in `Foo` — DeepSpace, May 15 '18 at 14:05
That's why it's quoted, just to separate it from other type of reference in context of the question. — Anton Ovsyannikov, May 15 '18 at 14:06
@AntonOvsyannikov: then perhaps don't call it a weak reference; just say you want to exclude something in the copy? — Martijn Pieters, May 15 '18 at 14:10

Martijn Pieters · Accepted Answer · 2018-05-15T14:13:27.307

Your class can implement a __deepcopy__ method to control how it is copied. From the copy module documentation:

In order for a class to define its own copy implementation, it can define special methods __copy__() and __deepcopy__(). The former is called to implement the shallow copy operation; no additional arguments are passed. The latter is called to implement the deep copy operation; it is passed one argument, the memo dictionary. If the __deepcopy__() implementation needs to make a deep copy of a component, it should call the deepcopy() function with the component as first argument and the memo dictionary as second argument.

Simply return a new instance of your class, with the reference you don't want to be deep-copied just taken across as-is. Use the deepcopy() function to copy other objects:

from copy import deepcopy

class Foo:
    def __init__(self, content, linked_to):
        self.content = content
        self.linked_to = linked_to

    def __deepcopy__(self, memo):
        # create a copy with self.linked_to *not copied*, just referenced.
        return Foo(deepcopy(self.content, memo), self.linked_to)

Demo:

>>> a1 = Foo([[1, 2], [3, 4]], None)
>>> a2 = Foo([[5, 6], [7, 8]], a1)
>>> a3 = deepcopy(a2)
>>> a3.linked_to.content.append([9, 10])  # still linked to a1
>>> a1.content
[[1, 2], [3, 4], [9, 10]]
>>> a1 is a3.linked_to
True
>>> a2.content is a3.content  # content is no longer shared
False

how does the memo parameter fall into this? I'm struggling to get my head around how to do this for an arbritrary number of properties I want linked and not linked — Rugnir, Nov 04 '21 at 00:05
@Rugnir you need to pass it along to other `deepcopy()` calls, nothing else. For every property you need to copy for the return value, call `deepcopy(value_to_copy, memo)`. — Martijn Pieters, Jan 19 '22 at 00:39

How to exclude specific references from deepcopy?

1 Answers1

Linked