如何在python中将展平表转换为嵌套(分层)json?

时间:2017-10-28 09:41:16

标签: python json nested flatten hierarchical

我有一张平坦的桌子。我想将其转换为嵌套的json格式。

database     schema table_name          col_name

exploration  dbo    TestTable           name
exploration  dbo    TestTable           last_name
exploration  dbo    Table_1             name
exploration  dbo    Table_1             d_id
exploration  dbo    tblMonitorChange    EventType
exploration  dbo    tblMonitorChange    SchemaName
exploration  dbo    tblMonitorChange    ObjectName
exploration  dbo    tblMonitorChange    ObjectType
exploration  dbo    tblMonitorChange    EventDate
exploration  dbo    tblMonitorChange    SystemUser
exploration  dbo    tblMonitorChange    CurrentUser
exploration  dbo    tblMonitorChange    OriginalUser
ReportServer dbo    Users               UserID
ReportServer dbo    Users               Sid
ReportServer dbo    Users               UserType
ReportServer dbo    Users               AuthType
ReportServer dbo    Users               UserName
ReportServer dbo    Users               ServiceToken
ReportServer dbo    Users               Setting

我正在寻找一般解决方案,而不是硬编码列名。

任何帮助将不胜感激。

1 个答案:

答案 0 :(得分:0)

我递归地解决了这个问题。这个功能测试和工作:

def table_to_json(model):
    print('-------------------------------------------------------------')
    doc = {};
    col_names = list(model.columns)
    grouped = model.groupby(col_names[0])[col_names[1]]
    values = grouped.apply(lambda x: set(x.tolist()))
    a = values.shape
    if(len(col_names)==2):
        return dict(values)
    keys = list(grouped.groups.keys())
    for k in keys:
        doc.update({k:table_to_json(model[model[col_names[0]] == k][col_names[1:]])})

    return doc

它将数据帧(在我的第一篇文章中检查)作为输入,并返回字典文档作为输出。这是输出的一部分:

{'exploration': {'dbo': {'Table_1': {'d_id', 'name'},
   'TestTable': {'last_name', 'name'},
   'tblMonitorChange': {'CurrentUser',
    'EventDate',
    'EventType',
    'ObjectName',
    'ObjectType',
    'OriginalUser',
    'SchemaName',
    'SystemUser'}}},
'ReportServer': {'dbo': {'ActiveSubscriptions': {'ActiveID',
    'SubscriptionID',
    'TotalFailures',
    'TotalNotifications',
    'TotalSuccesses'},
    'Batch': {'Action',
    'AddedOn',
    'BatchID',
    'BoolParam',
    'Content',
    'Item',
    'Param',
    'Parent',
    'Properties'}}}}
相关问题