[Python知识库]python对dataframe中series的json格式解析

方法1：如果df里只有一列json格式，可以保存为txt，然后再删掉列名，在进行处理。

import pandas as pd

result = []
with open(r"C:\Users\Administrator\Desktop\json处理.txt") as f:
    for line in f:
        print(line) # txt样子
        line = line.replace('""','"')
        line = line.replace('"{','{')
        line = line.replace('}"','}')
#         print(line) # 处理后的line
        dict_o = eval(line.strip()) #eval：将字符串字典转换为字典，strip：将某行字的两端空格去掉
        result.append(dict_o)
df = pd.DataFrame(result)
df

txt 内容

{""imei"":""5gy5"",""deviceid"":""c5ty7890h"",""accid"":""null""}

{"imei":"5gy5","deviceid":"c5ty7890h","accid":"null"}

{""imei"":""5rrr5"",""createtime"":""c5ty7890h"",""accid"":""null""}

{"imei":"5rrr5","createtime":"c5ty7890h","accid":"null"}

{""appid"":""5eeey5"",""deviceid"":""c5ty7890h"",""createtime"":""c5ty7890h"",""accid"":""null""}

{"appid":"5eeey5","deviceid":"c5ty7890h","createtime":"c5ty7890h","accid":"null"}

{""imei"":""null"",""deviceid"":""c544ty7890h"",""accid"":""null""}

{"imei":"null","deviceid":"c544ty7890h","accid":"null"}

结果输出：

方法2

import json
import pandas as pd
df = pd.read_excel(r"C:\Users\Administrator\Desktop\json处理.xlsx")
print(df)
data = df.to_dict(orient='records')
for _ in data:
    _.update(json.loads(_['json']))
    del _['json']

df1 = pd.DataFrame(data)
df1

原表

    城市                                               json
0   重庆  {"imei":"5gy5","deviceid":"c5ty7890h","accid":...
1   北京  {"imei":"5gy5","deviceid":"c5ty7890h","accid":...
2   上海  {"imei":"5rrr5","createtime":"c5ty7890h","acci...
3   广州  {"imei":"5rrr5","createtime":"c5ty7890h","acci...
4   深圳  {"appid":"5eeey5","deviceid":"c5ty7890h","crea...
5   地球  {"appid":"5eeey5","deviceid":"c5ty7890h","crea...
6   火星  {"imei":"null","deviceid":"c544ty7890h","accid...
7  太阳星  {"imei":"null","deviceid":"c544ty7890h","accid...

结果输出：

方法3，取json的固定字段

import json
import pandas as pd
df = pd.read_excel(r"C:\Users\Administrator\Desktop\json处理.xlsx")
print(df)
tmp=[]
for line in df.json.values:
#     d=eval(line)[0] 
    d=eval(line)
    tmp.append(d.get('imei') or d.get('deviceid'))
    # tmp.append(d.get('xx')['yy']) #如果是嵌套字典，就用此方法
df['结果列']=tmp
df

原表

    城市                                               json
0   重庆  {"imei":"5gy5","deviceid":"c5ty7890h","accid":...
1   北京  {"imei":"5gy5","deviceid":"c5ty7890h","accid":...
2   上海  {"imei":"5rrr5","createtime":"c5ty7890h","acci...
3   广州  {"imei":"5rrr5","createtime":"c5ty7890h","acci...
4   深圳  {"appid":"5eeey5","deviceid":"c5ty7890h","crea...
5   地球  {"appid":"5eeey5","deviceid":"c5ty7890h","crea...
6   火星  {"imei":"null","deviceid":"c544ty7890h","accid...
7  太阳星  {"imei":"null","deviceid":"c544ty7890h","accid...

结果输出：