简体   繁体   中英

Python pickling keep object identity

Is there any way to preserve the identity of a pickled object, ie have the below print True :

import pickle

class Foo:
    pass

x = Foo()
print(x is pickle.loads(pickle.dumps(x)))          #False

I am using cPickle and cpython 3.x on a Linux box, don't need something that's portable.

yes, it is possible; You'll need to include the "identity" in the pickled result some how; the most natural being to use __getnewargs__ and have a __new__ method return the existing, cached instance in that case.

import uuid
import weakref


class Foo(object):
    ident_cache = weakref.WeakValueDictionary()

    def __new__(cls, identity=None, **kwargs):
        if identity is None:
            identity = uuid.uuid1()
        try:
            self = cls.ident_cache[identity]
        except KeyError:
            self = super(Foo, cls).__new__(cls)
            self.__identity = identity
            self.__init__(**kwargs)
            cls.ident_cache[identity] = self
        return self

    def __getnewargs__(self):
        return (self.__identity,)

    def __init__(self, foo):
        self.foo = foo
>>> import pickle
>>> a = Foo(foo=1)
>>> b = pickle.loads(pickle.dumps(a, pickle.HIGHEST_PROTOCOL))
>>> a is b
True

The important note is that you must use protocol version 2 (or higher, hypothetically); because otherwise, __new__ is never called. This is only a concern on pickle.dumps , loads doesn't care.

import pickle

class Foo:
    _id_counter = 0
    def __init__(self):
        self._id = Foo._id_counter
        Foo._id_counter += 1

x = Foo()
print(x._id==pickle.loads(pickle.dumps(x))._id)     # True

The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM