如何在 Python 中展平其中一列包含 json 对象的数据框

假设json_object的每个值都是dict，也可以使用以下方法:

df = pd.DataFrame(
    data = {
        "customer_id": ["A101", "A102"],
        "date": ["2022-06-21", "2022-06-22"],
        "json_object": [{'name': 'james','age':55, 'hobby':'pubg'}, {'name': 'tarzan','status':'single'}]
    }
)
df["json_object"] = df["json_object"].map(lambda x: [[i, x[i]] for i in x])
df = df.explode(column="json_object")
df.json_object = df.json_object.str[0].astype(str) + ": " + df.json_object.str[1].astype(str) 
df

------------------------------------------
    customer_id  date        json_object
0   A101         2022-06-21  name: james
0   A101         2022-06-21  age: 55
0   A101         2022-06-21  hobby: pubg
1   A102         2022-06-22  name: tarzan
1   A102         2022-06-22  status: single
------------------------------------------

编辑

因为您将数据框更改为

df = pd.DataFrame(
    data = {
        "customer_id": ["A101", "A102"],
        "date": ["2022-06-21", "2022-06-22"],
        "json_object": [{'name': ['james'],'age':[55], 'hobby':['pubg']}, {'name': ['tarzan'],'status':['single']}]
    }
)

我的代码必须调整如下:

df = pd.DataFrame(
    data = {
        "customer_id": ["A101", "A102"],
        "date": ["2022-06-21", "2022-06-22"],
        "json_object": [{'name': ['james'],'age':[55], 'hobby':['pubg']}, {'name': ['tarzan'],'status':['single']}]
    }
)
df["json_object"] = df["json_object"].map(lambda x: [[i, x[i][0]] for i in x])
df = df.explode(column="json_object")
df.json_object = df.json_object.str[0].astype(str) + ": " + df.json_object.str[1].astype(str) 
df

如果包含空列表，只需在lambda函数中添加if-else条件即可.注意，我还在下一个代码提取中重命名了这些列.

df = pd.DataFrame(
    data = {
        "customer_id": ["A101", "A102"],
        "date": ["2022-06-21", "2022-06-22"],
        "json_object": [{'name': ['james'],'age':[55], 'hobby':['pubg']}, {'name': ['tarzan'],'status':[]}]
    }
)
df["json_object"] = df["json_object"].map(lambda x: [[i, x[i][0]] if x[i] else [i, ""] for i in x])
df = df.rename(columns={"json_object": "attribute"}).explode(column="attribute")
df.attribute = df.attribute.str[0].astype(str) + ": " + df.attribute.str[1].astype(str)