json&pickle serialization

Posted by nels on Sun, 16 Jun 2019 01:03:05 +0200

1. Purpose

You need to serialize the data in memory, that is, when you write to a file, you can only write to a string or binary type.But if we want to serialize more complex data types like lists, dictionaries, or functions, we use json or pickle.

2. json serialization

1. dumps serialization and loads deserialization

dumps converts data types to strings

import json

info = {
    'name': 'The Count of Monte Cristo',
    'type': 'Movie'
}

data = json.dumps(info)
print(data)
print(type(data))

# output
{"name": "The Count of Monte Cristo", "type": "Movie"}
<class 'str'>

loads converts strings to data types

import json

get_info = json.loads(data)
print(get_info['name'])
print(get_info)
print(type(get_info))

#output
The Count of Monte Cristo
{'name': 'The Count of Monte Cristo', 'type': 'Movie'}
<class 'dict'>

2.dump serialization and load deserialization

dump converts data types to strings and stores them in files

import json

info = {
    'name': 'The Count of Monte Cristo',
    'type': 'Movie'
}

with open("test.txt", "w", encoding="utf-8") as f:
    json.dump(info, f)  # The first parameter is an in-memory data object, and the second parameter is a file handle

#Write content to file
{"name": "The Count of Monte Cristo", "type": "Movie"}

load converts a file open from a string to a data type

import json


with open("test.txt", "r", encoding="utf-8") as f:
    data_from_file = json.load(f)

print(data_from_file['name'])
print(data_from_file)
print(type(data_from_file))

#output
The Count of Monte Cristo
{'name': 'The Count of Monte Cristo', 'type': 'Movie'}
<class 'dict'>

3.json serializes a function

import json

def test(name):
    print("hello,{}".format(name))

info = {
    'name': 'The Count of Monte Cristo',
    'type': 'Movie',
    'func': test
}

data = json.dumps(info)

#output
 File "G:/python/untitled/study6/json&pickle Modular.py", line 22, in <module>
    data = json.dumps(info)
  File "G:\python\install\lib\json\__init__.py", line 230, in dumps
    return _default_encoder.encode(obj)
  File "G:\python\install\lib\json\encoder.py", line 198, in encode
    chunks = self.iterencode(o, _one_shot=True)
  File "G:\python\install\lib\json\encoder.py", line 256, in iterencode
    return _iterencode(o, 0)
  File "G:\python\install\lib\json\encoder.py", line 179, in default
    raise TypeError(repr(o) + " is not JSON serializable")
TypeError: <function test at 0x0000021B13C57F28> is not JSON serializable

1. json can only handle simple data types, such as dictionaries, lists, strings, etc. it cannot handle complex data types such as functions.

2. JSON is common to all languages and all languages support json. If we need python to interact with data in other languages, use JSON format

3. Pile serialization

The use of a pickle is the same as above, but the serialized data type of the pickle is binary and the pickle can only be used in python.

1.dumps && loads

import pickle


def test(name):
    print("hello,{}".format(name))

info = {
    'name': 'The Count of Monte Cristo',
    'type': 'Movie',
    'func': test
}

data = pickle.dumps(info)
print(data)
print(type(data))

#output
b'\x80\x03}q\x00(X\x04\x00\x00\x00nameq\x01X\x19\x00\x00\x00The Count of Monte Cristoq\x02X\x04\x00\x00\x00typeq\x03X\x05\x00\x00\x00Movieq\x04X\x04\x00\x00\x00funcq\x05c__main__\ntest\nq\x06u.'

<class 'bytes'>

import pickle

get_data = pickle.loads(data)
get_data['func']('cat')
print(get_data)

#output
hello,cat
{'name': 'The Count of Monte Cristo', 'type': 'Movie', 'func': <function test at 0x00000235350A7F28>}

2. dump && load

import pickle


def test(name):
    print("hello,{}".format(name))

info = {
    'name': 'The Count of Monte Cristo',
    'type': 'Movie',
    'func': test
}

with open('test.txt', 'wb') as f:
    pickle.dump(info, f)

# Write to the contents of the test.txt file

�}q (X   typeqX   MovieqX   funcqc__main__
test
qX   nameqX   The Count of Monte Cristoqu.

import pickle

with open('test.txt', 'rb') as f:
    get_data = pickle.load(f)
print(get_data)

# output

{'name': 'The Count of Monte Cristo', 'func': <function test at 0x000001BA2AB4D510>, 'type': 'Movie'}

Summary:

The json value supports simple data types, and the pickle supports all data types.
pickle can only support python's own serialization and deserialization, not for data interaction with other languages, but json can.
Pile serializes the entire data object, so when you deserialize a function, the logic in the body of the function changes, following the body of the heart.

Topics: Python JSON encoding

Programmer Think