python 中 pickle 模块学习笔记

1. 概要
2. 和json 模块的比较
3. 常用api说明
4. 具体应用举例
5. 总结

6. 参考文档 pickle模块

   一. 概要

   用这个模块 可以创建Python对象的可移植序列化表示。
   Create portable serialized representations of Python objects.

   二.pickle 和json 模块对比

   1. json 模块实现序列化方式 一般是 unicode text ,而且序列化后, 人是可以看懂的. 但是 pickle 序列化方式是二进制的, 人是看不懂的.
   2. json 序列化更加广泛,并且json 是可以互操作的. pickle 仅是Python 独有的一种方式.
   3. json 可以序列化Python内置类型的子类. 但是我如果要序列化 自己写的类,显得无能为力. 但是pickle 模块是可以的,可以序列化 自己写的类,同时可以序列化Python内置类型的子类型.

There are fundamental differences between the pickle protocols and JSON (JavaScript Object Notation):

JSON is a text serialization format (it outputs unicode text, although most of the time it is then encoded to utf-8), while pickle is a binary serialization format;
JSON is human-readable, while pickle is not;
JSON is interoperable and widely used outside of the Python ecosystem, while pickle is Python-specific;
JSON, by default, can only represent a subset of the Python built-in types, and no custom classes; pickle can represent an extremely large number of Python types (many of them automatically, by clever usage of Python’s introspection facilities; complex cases can be tackled by implementing specific object APIs).

1看一个小例子

import pickle
import json

if __name__ == '__main__':
    d1 = dict(zip('frank', range(5)))
    print(d1)

    json_str = json.dumps(d1)
    pickle_str = pickle.dumps(d1)

    print(f'json_str: {json_str}')
    print(f'pickle_str: {pickle_str}')

结果如下:

{'f': 0, 'r': 1, 'a': 2, 'n': 3, 'k': 4}
json_str: {"f": 0, "r": 1, "a": 2, "n": 3, "k": 4}
pickle_str: b'\x80\x03}q\x00(X\x01\x00\x00\x00fq\x01K\x00X\x01\x00\x00\x00rq\x02K\x01X\x01\x00\x00\x00aq\x03K\x02X\x01\x00\x00\x00nq\x04K\x03X\x01\x00\x00\x00kq\x05K\x04u.'

Process finished with exit code 0

可以看出来,json 序列化后,是人类能够看懂的.而pickle 模块序列化后,就看不懂了,因为是二进制的.

在看一个例子

#!/usr/bin/env python3
# -*- coding: UTF-8 -*-
""" @author: Frank @contact: frank.chang@shoufuyou.com @file: test_pickle.py @time: 2018/7/22 上午9:06 """

import pickle
import json


class Person:
    __tablename__ = 'person'

    table_flag = 'online'

    def __init__(self, name):
        self.name = name

    @classmethod
    def pickup(cls, *args, **kwargs):
        print('pickup() is running.')
        kwargs.update({"name": "frank", "hobby": "swim"})
        return kwargs




def test_pickle():
    # 序列化类
    p1 = pickle.dumps(Person)

    # 反序列化
    P1 = pickle.loads(p1)
    # 打印Person类
    print(P1)
    p2 = P1('frank')
    print(p2.pickup())


def test_json():
    p1 = json.dumps(Person)

    P1 = json.loads(p1)
    print(P1)
    p2 = P1('frank')
    print(p2.pickup())


if __name__ == '__main__':
    # test_pickle()

    test_json()

报错如下:

TypeError: Object of type 'type' is not JSON serializable

这里就是 type 不可以json 序列化的.

而用 test_pickle() 是可以的.
结果如下:

<class '__main__.Person'>
pickup() is running.
{'name': 'frank', 'hobby': 'swim'}

##### 三. 常用api说明

提供了常用序列化,和反序列化的接口
dumps dump 前一个返回时一个bytes 对象 , 后一个直接序列化到文件里面

loads load 前一个 从二进制bytes对象读取对象, 后一个 从文件中读取对象

pickle.dump(obj, file, protocol=None, *, fix_imports=True)

pickle.dumps(obj, protocol=None, *, fix_imports=True)

pickle.load(file, *, fix_imports=True, encoding=”ASCII”, errors=”strict”)

pickle.loads(bytes_object, *, fix_imports=True, encoding=”ASCII”, errors=”strict”)

如果要有更多的空值, 可以使用下面的两个类 来定制你的 序列化对象
The pickle module exports two classes, Pickler and Unpickler:

如果要对序列化和反序列化进行更多控制，可以分别创建Pickler或Unpickler对象。

pickle 模块定义的异常
The pickle module defines three exceptions:

exception pickle.PickleError

exception pickle.PicklingError

exception pickle.UnpicklingError

来看一个例子

import pickle


class Person:
    __tablename__ = 'person'

    table_flag = 'online'

    def __init__(self, name):
        self.name = name

    @classmethod
    def pickup(cls, *args, **kwargs):
        print('pickup() is running.')
        # 所有的参数直接返回,不做任何处理.

        kwargs.update({"name": "frank", "hobby": "swim"})
        return kwargs

    @classmethod
    def extract(cls, value='frank'):
        """获取写数据库必要数据 :param value: pickup 方法的返回值 :param context: pickup 方法的入参 :return: """
        print('extract() is running.')
        return value


class Serialization:

    def __init__(self, obj):
        self.myclass = obj

    def serialize(self):
        with open('pickle.txt', 'wb+') as f:
            # 写入 序列化到文件
            pickle.dump(self.myclass, f)

    def deserialize(self):
        # 反序列化 从文件反序列化
        with open('pickle.txt', 'rb') as f:
            # 读取
            data = pickle.load(f)
        return data



if __name__ == '__main__':
    ser = Serialization(Person)

    ser.serialize()

    person = ser.deserialize()

    print(f'person.table_flag: {person.table_flag}')

    print(person.pickup())
    print(person.extract())

结果如下:

person.table_flag: online
pickup() is running.
{'name': 'frank', 'hobby': 'swim'}
extract() is running.
frank

这个例子就是把类序列化到文件里面, 之后再从文件中读出来.

四. 具体应用举例

这个模块具体有什么用呢? 比如 有一个系统需要动态加载类, (我的意思是类是通过代码生成的,然后要把这个类加载到内存里面)
但是有一天我担心,如果程序突然有意外的bug ,或者其他的情况崩溃了, 而之前加载的类,就会消失了,一旦重启了系统,所有动态生成的类就会消失了, 所以pickle 模块就给我提供非常好用的方法. 可以把类序列化写到文件,或者序列化到二进制bytes 对象. 之后如果系统重启后,我重新 反序列把类读取到内存里面,完成反序列化.

#!/usr/bin/env python3
# -*- coding: UTF-8 -*-
""" @author: Frank @contact: frank.chang@shoufuyou.com @file: serialization.py @time: 2018/7/22 上午12:18 """

import pickle


class Person:
    __tablename__ = 'person'

    table_flag = 'online'

    def __init__(self, name):
        self.name = name

    @classmethod
    def pickup(cls, *args, **kwargs):
        print('pickup() is running.')
        # 所有的参数直接返回,不做任何处理.

        kwargs.update({"name": "frank", "hobby": "swim"})
        return kwargs

    @classmethod
    def extract(cls, value='frank'):
        """ :param value: pickup 方法的返回值 :return: """
        print('extract() is running.')
        return value


class Serialization:

    def __init__(self):
        self.myclasses = []

    def resigester(self, obj):
        self.myclasses.append(obj)

    def serialize(self):
        # 写入 序列化
        pickle_strings = []
        for myclass in self.myclasses:
            pickle_string = pickle.dumps(myclass)

            pickle_strings.append(pickle_string)

        return pickle_strings

    @staticmethod
    def deserialize(bytes_object):
        # 反序列化 从文件反序列化
        return pickle.loads(bytes_object)


if __name__ == '__main__':

    serialization = Serialization()

    serialization.resigester(Person)

    # 序列化Person 类
    strings = serialization.serialize()
    # 打印 序列化的结果
    print(strings)

    for bytes_obj in strings:
        # 反序列,得到Person 类
        P = serialization.deserialize(bytes_obj)
        print(P)

        # 构造p1 对象
        p1 = P('frank')
        print(p1.pickup())

结果如下:

[b'\x80\x03c__main__\nPerson\nq\x00.']
<class '__main__.Person'>
pickup() is running.
{'name': 'frank', 'hobby': 'swim'}

通过Serialization 序列化 Person类,之后有把他反序列出来, 完成序列化, 与反序列化操作.

五.参考文档

本文简单介绍了pickle模块的常见用法,常用api , 比较了与pickle 模块的不同. 如果需要特殊定制序列化, 可以使用 接口提供的那两个类.Pickler and Unpickler 这两个类更多的参考官方文档, 一般用的比较少.

六.参考文档

https://docs.python.org/3/library/pickle.html