Python:如何用匹配的键将字典列炸成列？

Question

我在“熊猫数据”中有一篇专栏文章，它的结构如下(参见示例)。我想我在一个列中有一个嵌套的字典，我希望每个键都有自己的列。我希望所有匹配的键都是同一列。运行示例了解更多详细信息

import pandas as pd
import numpy as np 

data = ["{'age': 59, 'gender': 'Female','pain': 'No', 'date': '2022-09-29'}",
       "{'gender': 'Male', 'date': '2022-10-11'}",
       "{'age': 18, 'date': '2022-10-11', 'pain': 'No'}",
       ]

original_df = pd.DataFrame(data, columns=['test'])
original_df.head()

我想炸开dataframe，使它具有以下结构：

data = [[59,'Female','No','2022-09-29'], [np.nan,'Male',np.nan,'2022-10-11'], [18,np.nan,'2022-10-11','No']]
  
# Create the pandas DataFrame
desired_df = pd.DataFrame(data, columns=['age','gender','date','pain'])
desired_df.head()

Answer 1

使用json.normalize

from ast import literal_eval

import pandas as pd


original_df["test"] = original_df["test"].apply(literal_eval)
df = original_df.join(pd.json_normalize(original_df.pop("test")))
print(df)

    age  gender pain        date
0  59.0  Female   No  2022-09-29
1   NaN    Male  NaN  2022-10-11
2  18.0     NaN   No  2022-10-11