前言：

Vue框架：从项目学Vue
OJ算法系列：神机百炼 - 算法详解
Linux操作系统：风后奇门 - linux
C++11：通天箓 - C++11

应用场景：

读取一个json文件，利用其中某些键值对来构建一个新的json文件

• 如获取下列json文件中某几行的键值对，再构建一个json文件

{"packs": [{"_id": "1","time": 123,"category": 0,"current_info": {"tag": ["timestamp"],"fps": 60.0,"time_stamp": 414.514,"name": "test"},"content": {"core": "service","status": 0,"extraction": "client"}}]
}

基本工具：

文件操作：

打开文件：

• open(‘文件路径’, ‘打开方式’)：
  1. w：覆盖写
  2. r：只读
  3. wb：二进制写
  4. rb：二进制读

写文件：

• write(字符串)：
  字符串中转义字符：
  1. \r：切换到当前行的行首列
  2. \n：切换到下一行的同一列

file = open('Json文件地址', 'w')
lines = ''
for line in file:#line自动省略\r\nlines += line

读文件：

• open()返回一个可迭代对象：

file = open('Json文件地址', 'r')
lines = ''
for line in file:#line自动省略\r\nlines += line

关闭文件并刷新缓冲区：

• close：

file.close()

• close()后会自动刷新缓冲区，但有时你需要在关闭前刷新它，这时就可以使用 flush() 方法：

file.flush()

Json字符串和字典转换：

json.loads()：

• 用于将 字符串 加载为 字典

file = open('Json文件地址', 'r')
lines = ''
for line in file:lines += line
dic = json.loads(lines)
print(dic)

json.dumps():

• 用于将 字典 写入到 字符串

dic = {'packs': [{'_id': '1', 'time': 123, 'category': 0, 'current_info': {'tag': ['timestamp'], 'fps': 60.0, 'time_stamp': 414.514, 'name': 'test'}, 'content': {'core': 'service', 'status': 0, 'extraction': 'client'}}]}json_str = json.dumps(dic)print(json_str)

Json文件和字典转化：

json.load()：

• 用于将文件 加载为 字典

dictionary = json.load(open('原本的Json文件地址', 'r'))
print(dictionary)

json.dump():

• 用于将字典 写入到 文件，创建文件使用w模式打开

json.dump(ret, open("新的Json文件地址","w"))

集成上述API为函数：

def func(dictionary,*direct_key, **indirect_key):ret = {}for i in direct_key:ret[i] = dictionary[i]for i in indirect_key:ret[i] = {}print(type(ret), type(ret[i]), i)for j in indirect_key[i]:print(j)ret[i][j] = dictionary[i][j]return retret = func(json.load('原来的Json文件地址'), '_id', 'end_time', 'frames', ext=['core_cm', 'extraction_cm'], statistics=['parsing_begin', 'parsing_end'])json.dump(ret, open("新的Json文件地址","w"))

自己开发一个json.dump：

• 简化版，只能接收json最多二重字典，且最终value为str, int, 或float

import os
import json
def isIntOrFloatSeriously(number):result = Falsetry:n = float(number)if str(number).count('.') != 0 or (n.is_integer() and str(number).count('.') == 0):result = Trueexcept:result = Falsereturn resultdef func(dictionary,*direct_key, **indirect_key):amount = len(direct_key) + len(indirect_key)count = 1ret = ''for i in direct_key:ret += f"\"{i}\": "if isIntOrFloatSeriously(dictionary[i]) == True :ret += str(dictionary[i])else:ret += "\""+dictionary[i]+"\""if count < amount:ret += ","ret += "\r\n"count = count + 1for i in indirect_key:ret += f"\"{i}\": "+'{\r\n'inner_count = 1inner_amount = len(indirect_key[i])print(inner_amount, indirect_key[i])for j in indirect_key[i]:ret += f"\"{j}\": "if isIntOrFloatSeriously(dictionary[i][j]) == True :ret += str(dictionary[i][j])else:ret += "\""+dictionary[i][j]+"\""if inner_count < inner_amount:ret += ","ret += "\r\n"inner_count = inner_count + 1ret += "}"if count < amount:ret += ","ret += "\r\n"count = count + 1return retfile = open('原本的Json文件地址')
lines = ''
for line in file:lines += line
dic = json.loads(lines)ret = func(dic["packs"][0], '_id', 'end_time', 'frames', ext=['core_cm', 'extraction_cm'], statistics=['parsing_begin', 'parsing_end'])file = open('新建的Json文件地址', 'w')
file.write('{\r\n')
file.write(ret)
file.write('}\r\n')
file.close()

扩展：

• Yaml文件格式
• Yaml文件 转化为 字典
  1. yaml.safe_load()：不省略float的.0
  2. yaml.load()：省略float的.0