Python 3.x CSV到嵌套JSON

我的CSV是这样的：

csId,lut,seqId,lvlId,lvlTyp,accSt,enrlDt,ptnrName,ptnrIds
 27768303,1561939200,1,G,GAR,10,06-06-2018,Chase,12345
 27768303,1561939200,1,G,GAR,10,06-06-2018,Chase,98765
 27768303,1561939200,1,G,GAR,10,06-06-2018,fliggy,67890
 68537125,1562025600,2,S,SAR,20,11-12-2014,fliggy,98696

但我得到的是：

[{"accSt": "10", 
   "csId": 27768303,
    "enrlDt": "06-06-2018",
    "lut": 1561939200,
    "lvlId": "G",
    "lvlTyp": "GAR",
     "ptnrlst": "ptnrName":"Chase","ptnrIds":12345},
     "seqId": 1,
     "type": "mber"},
   {"accSt": "10",
    "csId": 27768303,
    "enrlDt": "06-06-2018",
    "lut": 1561939200,
    "lvlId": "G",
    "lvlTyp": "GAR",
    "ptnrlst": {"ptnrName":"Chase","ptnrIds":98765},
    "seqId": 1,
    "type": "mber"},
  { "accSt": "10",
    "csId": 27768303,
    "enrlDt": "06-06-2018",
    "lut": 1561939200,
    "lvlId": "G",
    "lvlTyp": "GAR",
    "ptnrlst": {"ptnrName":"fliggy","ptnrIds":67890},
    "seqId": 1,
    "type": "mber"},
 { "accSt": "20",
  "csId": 68537125,
  "enrlDt": "11-12-2014",
   "lut": 1562025600,
   "lvlId": "S",
   "lvlTyp": "SAR",
   "ptnrlst": {"ptnrName":"Chase","ptnrIds":98696},
   "seqId": 2,
   "type": "mber"}]

我尝试使用以下代码：

from csv import DictReader
from itertools import groupby
from pprint import pprint
import fileinput
import time


 def check_for_null_new(allvars):
     first_split =  allvars.split(',')
     ret_val=""
     loop_cnt=1
     for second_split in first_split:
             individual_split = second_split.split(':')
             if not individual_split[1]:
                 pass
             else:
                 if loop_cnt == 1:
                    if individual_split[1].isnumeric():
                        ret_val=(individual_split[0]+":"+individual_split[1])
                    else:
                        ret_val= 
                       (individual_split[0]+":'"+individual_split[1]+"'")
                    else:
                          if individual_split[1].isnumeric():
                               ret_val=ret_val +','+   
                            (individual_split[0]+":"+individual_split[1])
                          else:
                                ret_val=ret_val +','+ 
                       (individual_split[0]+":'"+individual_split[1]+"'")
                loop_cnt = loop_cnt + 1


    return (ret_val)

start_time = time.time()
with open('member.csv',encoding='utf-8-sig') as csvfile:
    r1 = DictReader(csvfile, skipinitialspace=True)
    data = [dict(d) for d in r1]
    groups = []
    uniquekeys = []


    for k, g in groupby(data, lambda r: ( r['csId'], r['lut'], r['seqId'],r['lvlId'], r['lvlTyp'], r['accSt'], r['enrlDt'],r['ptnrName'],r['ptnrIds'])):

        groups.append({ "type": "mber",
                    "csId": int(k[0]),
                    "lut": int(k[1]),
                    "seqId": int(k[2]),
                    "lvlId": k[3],
                    "lvlTyp": k[4],
                    "accSt": k[5],
                    "enrlDt": k[6],
                    "ptnrlst" : {check_for_null_new("'ptnrName':"+k[7]+",'ptnrIds':"+ k[8])} 
                    })
    uniquekeys.append(g)

with open('member.json', 'wt') as out:
    pprint(groups, stream=out)

with fileinput.FileInput('member.json', inplace=True, backup='.bak') as file:
    for line in file:
        print(line.replace("\"", "").replace("'", "\""), end='')

end_time = time.time()
print("CSV to JSON Completed for Member in %s seconds " % (end_time - start_time))

我的预期产出是：

[
{
  "type": "mber",
  "csId": 27768303,
  "lut": 1561939200,
  "seqId":1,
  "lvlId": "G",
  "lvlTyp": "GAR",
  "accSt": "10",
  "enrlDt": "06-06-2018",
   "ptnrlst":[{"ptnrName":"Chase",
            "ptnrIds":["12345","98765"]
          },
          {"ptnrName":"fliggy",
            "ptnrIds":["67890"]
          }]
},
{
  "type": "mber",
  "csId": 68537125,
  "lut": 1562025600,
  "seqId":2,
  "lvlId": "S",
  "lvlTyp": "SAR",
  "accSt": "20",
  "enrlDt": "11-12-2014",
  "ptnrlst":[{"ptnrName":"chase","ptnrIds":["98696"]
              }]
}
]

---

特别回答以下问题：您在哪里读取一个

csv

文件，该文件输出一个包含嵌套对象的

json

文件，在您的例子中，似乎是

ptnrName

和

ptnrIds

初始化和读取文件应该是直截了当的，并考虑到您可以轻松地将整个文件加载到内存中

import csv
import itertools
from operator import itemgetter

with open('members.csv', 'r') as csvfile:
    all_ = list(csv.DictReader(csvfile, skipinitialspace=True))

当我们使用

itertools

groupby

函数时，这将是除上述嵌套对象之外的所有字段

keys = 'csId lut seqId lvlId lvlTyp accSt enrlDt'.split()
all_.sort(key = itemgetter(*(keys)))

下一部分是使用2个

groupby

函数创建嵌套操作

第一个

groupby

标识键以及键是否具有嵌套对象，第二个

groupby

根据

ptnrName

对键进行分组。把它们放在一起，你会得到：

import csv
import itertools
import json
from pprint import pprint
from operator import itemgetter

with open('members.csv', 'r') as csvfile:
    all_ = list(csv.DictReader(csvfile, skipinitialspace=True))

keys = 'csId lut seqId lvlId lvlTyp accSt enrlDt'.split() #list of keys
all_.sort(key = itemgetter(*(keys))) #inplace sort based on keys

ds = []

#1st groupby based on keys
for k, g in itertools.groupby(all_, key=lambda r: [r[i] for i in keys]):


    d = {key:value for key, value in zip(keep,k)} #create the default key,values 
    d['seqId'] = int(d['seqId'])

    for k1, g1 in itertools.groupby(g, key = lambda r: r['ptnrName']): #2nd groupby ptnrName

        array = [i['ptnrIds'] for i in g1] #array of multiple ptnrIds based on ptnrName

        #set default key ptnrlst to a list to store nested ptnrName, ptnrIds pairs
        d.setdefault('ptnrlst', []).append({'ptnrName':k1, 'ptnrIds':array})

    ds.append(d)

结果与预期一样，还要注意，在使用字典时，键没有排序

[{'accSt': '10',
  'csId': '27768303',
  'enrlDt': '06-06-2018',
  'lut': '1561939200',
  'lvlId': 'G',
  'lvlTyp': 'GAR',
  'ptnrlst': [{'ptnrIds': ['12345', '98765'], 'ptnrName': 'Chase'},
              {'ptnrIds': ['67890'], 'ptnrName': 'fliggy'}],
  'seqId': 1},
 {'accSt': '20',
  'csId': '68537125',
  'enrlDt': '11-12-2014',
  'lut': '1562025600',
  'lvlId': 'S',
  'lvlTyp': 'SAR',
  'ptnrlst': [{'ptnrIds': ['98696'], 'ptnrName': 'fliggy'}],
  'seqId': 2}]

最后转储到json：

with open('member.json', 'w') as jsonfile:
    json.dump(ds, jsonfile)

with open('member.json', 'r') as jsonfile:
    jload = json.load(jsonfile)

jload == ds

>>True

---

谢谢你，伯纳德。但是，所有的.sort（key=itemgetter（*（keep）））都会抛出“NameError:name'keep'未定义”。与语句d={key:value for key，value in zip（keep，k）}相同，键csId、lut和seqId必须是整数，但这里它们是stringsUpdated，key最初是keep，我添加了一行作为示例，在创建原始dict后将字段转换为

int

。没问题，编码愉快