将数据框中列中的 JSON/list 个字典拆分为 python 中的新行

Question

我是 Python 的新手，我试图找到答案，但似乎没有任何效果。当整个数据采用 JSON 格式

时，会提供大部分答案

通过PYODBC我使用下面的代码来检索数据

formula = """select  id, type, custbody_attachment_1 from transaction  """
lineitem = pd.read_sql_query(formula, cnxn)

它给了我类似下面的东西

+-------------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| Internal_ID | Type | Formula_Text                                                                                                                                                                                           |
+-------------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| 2895531     | Bill |                                                                                                                                                                                                        |
+-------------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| 3492009     | Bill | [{"FL":"https://.app.netsuite.com/core/media/media.nl?id=someLinkToTheFile0","NM":"someFileName0"}]                                                                                                    |
+-------------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+
| 3529162     | Bill | [{"FL":"5https://.app.netsuite.com/core/media/media.nl?id=someLinkToTheFile1","NM":"someFileName1"},{"FL":"https://.app.netsuite.com/core/media/media.nl?id=someLinkToTheFile2","NM":"someFileName2"}] |
+-------------+------+--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+

我需要这样的输出。（单元格中可能有超过 2 个链接。）

+-------------+------+---------------------------------------------------------------------+---------------+
| Internal_ID | Type | FL                                                                  | NM            |
+-------------+------+---------------------------------------------------------------------+---------------+
| 2895531     | Bill |                                                                     |               |
+-------------+------+---------------------------------------------------------------------+---------------+
| 3492009     | Bill | https://.app.netsuite.com/core/media/media.nl?id=someLinkToTheFile0 | someFileName0 |
+-------------+------+---------------------------------------------------------------------+---------------+
| 3529162     | Bill | https://.app.netsuite.com/core/media/media.nl?id=someLinkToTheFile1 | someFileName1 |
+-------------+------+---------------------------------------------------------------------+---------------+
| 3529162     | Bill | https://.app.netsuite.com/core/media/media.nl?id=someLinkToTheFile2 | someFileName2 |
+-------------+------+---------------------------------------------------------------------+---------------+

我试着玩 JSON 但问题一个接一个（因为对我来说它看起来像 JSON 数据）。最后我运行

print(lineitem['custbody_attachment_1'])

并在 Python 控制台中得到以下内容

999    [{"FL":"https://4811553.app.netsuite.com/core/...
Name: custbody_attachment_1, Length: 1000, dtype: object

所以，我不知道如何转换它以便创建新行

Answer 1

df = df.explode('Formula_Text')
df = pd.concat([df.drop(['Formula_Text'], axis=1), df['Formula_Text'].apply(pd.Series)], axis=1)
print(df)

将数据框中列中的 JSON/list 个字典拆分为 python 中的新行

split JSON/list of dictionaries in the column in dataframe to new rows in python

python

json

transform

object

dtype