将 csv 数据转换为 json 并添加新的字段名称“parsed_address”并基于地址类型值将放置在 parsed_address 字段中。 我可以创建 parsed_address 字段并将地址列放置在其下方，但地址字段应根据 address_type 放置。我在同一地址类型中获取特定 id 的所有地址，并在下一个地址类型中重复。

输入数据

"source_id"|"first_name"|"last_name"|"address_type"|"address_line_1"|"city"
"41614335"|Reinaldo|Tonkoski Jr.|Primary|Deh 211 Box 2222|Brookings|
"41614335"|Reinaldo|Tonkoski Jr.|home|"2409 10th St Apt 123"|Brookings
"07605348"|E|Christodoulou|Primary|"4D Ag Lavras st"|Kifissia
"07605348"|E|Christodoulou|home|"131 N Hamilton Dr Apt 308"|Beverly Hills

输出获取

[
   {
      "source_id":7605348,
      "first_name":"E",
      "last_name":"Christodoulou",
      "parsed_address":{
         "address_type":"Primary",
         "address":[
            {
               "address_line_1":"Deh 211 Box 2222",
               "city":"Brookings"
            },
            {
               "address_line_1":"4D Ag Lavras st",
               "city":"Kifissia"
            }
         ]
      }
   },
   {
      "source_id":41614335,
      "first_name":"Reinaldo",
      "last_name":"Tonkoski Jr.",
      "parsed_address":{
         "address_type":"home",
         "address":[
            {
               "address_line_1":"2409 10th St Apt 123",
               "city":"Brookings"
            },
            {
               "address_line_1":"131 N Hamilton Dr Apt 308",
               "city":"Beverly Hills"
            }
         ]
      }
   }
]

预期输出

[
   {
      "source_id":7605348,
      "first_name":"E",
      "last_name":"Christodoulou",
      "parsed_address":{
         "address_type":"Primary",
         "address":
            {
               "address_line_1":"Deh 211 Box 2222",
               "city":"Brookings"
            }
         "address_type":"home",
         "address" :
            {
               "address_line_1":"4D Ag Lavras st",
               "city":"Kifissia"
            }

TL;DR

from ast import literal_eval
from io import StringIO
from pprint import pprint

inputs = StringIO(""""source_id"|"first_name"|"last_name"|"address_type"|"address_line_1"|"city"
"41614335"|Reinaldo|Tonkoski Jr.|Primary|Deh 211 Box 2222|Brookings
"41614335"|Reinaldo|Tonkoski Jr.|home|"2409 10th St Apt 123"|Brookings
"07605348"|E|Christodoulou|Primary|"4D Ag Lavras st"|Kifissia
"07605348"|E|Christodoulou|home|"131 N Hamilton Dr Apt 308"|Beverly Hills""")


df = pd.read_csv(inputs, sep='|')

# Format the parsed_address to the dictionary format you want.
df['parsed_address'] = [{'address_type':row['address_type'], 'address':{'address_line_1':row['address_line_1'], 'city':row['city']}} 
                         for _, row in df.iterrows()]

lol = []
# Group by the source_id
for _, group_rows in df.groupby('source_id'):
    # Combine all parse_address for each group
    parsed_address = list(group_rows['parsed_address'])
    # Keeps the values of source_id, first and last name.
    group_rows = group_rows.iloc[0][['source_id', 'first_name', 'last_name']]
    # Set the parsed_address to the list you've combined.
    group_rows['parsed_address'] = parsed_address
    lol.append(group_rows)

js = literal_eval(pd.DataFrame(lol).to_json(orient='records'))

pprint(js)

[out]:

[{'first_name': 'E',
  'last_name': 'Christodoulou',
  'parsed_address': [{'address': {'address_line_1': '4D Ag Lavras st',
                                  'city': 'Kifissia'},
                      'address_type': 'Primary'},
                     {'address': {'address_line_1': '131 N Hamilton Dr Apt 308',
                                  'city': 'Beverly Hills'},
                      'address_type': 'home'}],
  'source_id': 7605348},
 {'first_name': 'Reinaldo',
  'last_name': 'Tonkoski Jr.',
  'parsed_address': [{'address': {'address_line_1': 'Deh 211 Box 2222',
                                  'city': 'Brookings'},
                      'address_type': 'Primary'},
                     {'address': {'address_line_1': '2409 10th St Apt 123',
                                  'city': 'Brookings'},
                      'address_type': 'home'}],
  'source_id': 41614335}]