如何将一列数据帧转换为具有 json 格式值的分布列（PYTHON）

Question

假设我有一个数据框值

例如 test_data.csv（包含以下数据）

effective_date,ds,id,id_type,e_data,create_id,create_timestamp
2021-07-26,am,27,a_id,"{""cup_id"": ""ffdsds"", ""rate"": ""B"", ""direct"": ""stable"", ""dl_tstmp"": ""2021-07-26 00:00:00"", ""inst_id"": 1213, ""src_p_tstmp"": ""2021-07-26 00:00:00"", ""inst_name"": ""abc corp""}",MA,2021-07-26 00:00:00
2021-07-26,am,24,a_id,"{""cup_id"": ""ererwe"", ""rate"": ""AB"", ""direct"": ""improvent"", ""dl_tstmp"": ""2021-07-26 00:00:00"", ""inst_id"": 66641, ""src_p_tstmp"": ""2021-07-26 00:00:00"", ""inst_name"": ""xyz corp""}",MA,2021-07-26 00:00:00
2021-07-27,am,22,a_id,"{""cup_id"": ""34kf3"", ""rate"": ""AA"", ""direct"": ""improvent"", ""dl_tstmp"": ""2021-07-26 00:00:00"", ""inst_id"": 6871, ""src_p_tstmp"": ""2021-07-26 00:00:00"", ""inst_name"": ""rimr corp""}",MA,2021-07-26 00:00:00
2021-07-27,am,32,a_id,"{""cup_id"": ""5gh23"", ""rate"": ""AAA"", ""direct"": ""downfall"", ""dl_tstmp"": ""2021-07-26 00:00:00"", ""inst_id"": 98795, ""src_p_tstmp"": ""2021-07-26 00:00:00"", ""inst_name"": ""prst corp""}",MA,2021-07-26 00:00:00


import pandas as pd
df = pd.read_csv("test_data.csv")

其中e_data列采用json_format而不是字典格式

我想将其分发为单独的列格式，因此预期输出如下

Answer 1

首先将字符串 dict 转换为真正的 dict：

from ast import literal_eval

df['e_data']=df['e_data'].map(literal_eval)

最后：

尝试使用join() + DataFrame() + tolist()和pop()删除“e_data”列：

df=df.join(pd.DataFrame(df.pop('e_data').tolist()))

要么

df=df.join(df['e_data'].apply(pd.Series)).drop('e_data',1)