Python 3.x 使用json_normalize从嵌套值构建表

当记录路径指向一列时，我在使用json_normalize时遇到了麻烦，其中有一个嵌套的dict，该dict随后包含一个列表。例如，见下文

鉴于以下情况：

list_of_dict = [
{
    'SCHOOL_NAME': 'SCHOOL_A',
    'STUDENTS': [
        {
            'STUDENT_NAME': 'JOHN',
            'STUDENT_ID': '1'
        },
        {
            'STUDENT_NAME': 'JANE',
            'STUDENT_ID': '2'
        },
    ]
},
{
    'SCHOOL_NAME': 'SCHOOL_B',
    'STUDENTS': [
        {
            'STUDENT_NAME': 'HENRY',
            'STUDENT_ID': '1'
        },
        {
            'STUDENT_NAME': 'MARK',
            'STUDENT_ID': '2'
        },
    ]
}]

我可以用电脑把它弄平

pd.json_normalize(data=list_of_dict, record_path='STUDENTS', meta=['SCHOOL_NAME'])[['SCHOOL_NAME', 'STUDENT_ID', 'STUDENT_NAME']]

要获得以下信息：

list_of_dict = [
{
    'SCHOOL_NAME': 'SCHOOL_A',
    'STUDENTS': [
        {
            'STUDENT_NAME': 'JOHN',
            'STUDENT_ID': '1'
        },
        {
            'STUDENT_NAME': 'JANE',
            'STUDENT_ID': '2'
        },
    ]
},
{
    'SCHOOL_NAME': 'SCHOOL_B',
    'STUDENTS': [
        {
            'STUDENT_NAME': 'HENRY',
            'STUDENT_ID': '1'
        },
        {
            'STUDENT_NAME': 'MARK',
            'STUDENT_ID': '2'
        },
    ]
}]

如果dict的列表结构如下，如何获得类似的输出格式： **注意增加了学生名单**

list_of_dict = [
{
    'SCHOOL_NAME': 'SCHOOL_A',
    'STUDENT_LIST':{
        'STUDENTS': [
            {
                'STUDENT_NAME': 'JOHN',
                'STUDENT_ID': '1'
            },
            {
                'STUDENT_NAME': 'JANE',
                'STUDENT_ID': '2'
            },
        ]
    }

},
{
    'SCHOOL_NAME': 'SCHOOL_B',
    'STUDENT_LIST': {
        'STUDENTS': [
            {
                'STUDENT_NAME': 'HENRY',
                'STUDENT_ID': '1'
            },
            {
                'STUDENT_NAME': 'MARK',
                'STUDENT_ID': '2'
            },
        ]
    }
}]

---

与

pop

一起使用

dict comprehension

：

# Just pop key `STUDENT_LIST` and your list_of_dict is back like before
In [680]: a = [{**x, **x.pop('STUDENT_LIST')} for x in list_of_dict]

# Now use `json_normalize`
In [684]: pd.json_normalize(a, record_path='STUDENTS', meta=['SCHOOL_NAME'])
Out[684]: 
  STUDENT_NAME STUDENT_ID SCHOOL_NAME
0         JOHN          1    SCHOOL_A
1         JANE          2    SCHOOL_A
2        HENRY          1    SCHOOL_B
3         MARK          2    SCHOOL_B