如何将 Python Dict 映射到 Big Query Schema

Question

我有一个带有一些嵌套值的字典，如下所示：

my_dict = {
    "id": 1,
    "name": "test",
    "system": "x",
    "date": "2015-07-27",
    "profile": {
        "location": "My City",
        "preferences": [
            {
                "code": "5",
                "description": "MyPreference",
            }
        ]
    },
    "logins": [
        "2015-07-27 07:01:03",
        "2015-07-27 08:27:41"
    ]
}

并且，我有一个 Big Query Table Schema，如下所示：

schema = {
    "fields": [
        {'name':'id', 'type':'INTEGER', 'mode':'REQUIRED'},
        {'name':'name', 'type':'STRING', 'mode':'REQUIRED'},
        {'name':'date', 'type':'TIMESTAMP', 'mode':'REQUIRED'},
        {'name':'profile', 'type':'RECORD', 'fields':[
            {'name':'location', 'type':'STRING', 'mode':'NULLABLE'},
            {'name':'preferences', 'type':'RECORD', 'mode':'REPEATED', 'fields':[
                {'name':'code', 'type':'STRING', 'mode':'NULLABLE'},
                {'name':'description', 'type':'STRING', 'mode':'NULLABLE'}
            ]},
        ]},
        {'name':'logins', 'type':'TIMESTAMP', 'mode':'REPEATED'}
    ]
}

我想遍历所有原始的 my_dict 并根据架构的结构构建一个新的 dict。换句话说，迭代模式并从原始 my_dict 中选取正确的值。

要构建这样的新字典（请注意，不复制模式中不存在的字段“系统”）：

new_dict = {
    "id": 1,
    "name": "test",
    "date": "2015-07-27",
    "profile": {
        "location": "My City",
        "preferences": [
            {
                "code": "5",
                "description": "MyPreference",
            }
        ]
    },
    "logins": [
        "2015-07-27 07:01:03",
        "2015-07-27 08:27:41"
    ]
}

如果没有嵌套字段迭代简单的 dict.items() 和复制值，可能会更容易，但是如何构建新的 dict 以递归方式访问原始 dict？

Answer 1

我已经构建了一个递归函数来做到这一点。我不确定这是否更好，但有效：

def map_dict_to_bq_schema(source_dict, schema, dest_dict):
    #iterate every field from current schema
    for field in schema['fields']:
        #only work in existant values
        if field['name'] in source_dict:
            #nested field
            if field['type'].lower()=='record' and 'fields' in field:
                #list
                if 'mode' in field and field['mode'].lower()=='repeated':
                    dest_dict[field['name']] = []
                    for item in source_dict[field['name']]:
                        new_item = {}
                        map_dict_to_bq_schema( item, field, new_item )
                        dest_dict[field['name']].append(new_item)
                #record
                else:
                    dest_dict[field['name']] = {} 
                    map_dict_to_bq_schema( source_dict[field['name']], field, dest_dict[field['name']] )
            #list
            elif 'mode' in field and field['mode'].lower()=='repeated':
                dest_dict[field['name']] = []
                for item in source_dict[field['name']]:
                    dest_dict[field['name']].append(item)
            #plain field
            else:
                dest_dict[field['name']]=source_dict[field['name']]

                format_value_bq(source_dict[field['name']], field['type'])

new_dict = {}
map_dict_to_bq_schema (my_dict, schema, new_dict)