Logo Questions Linux Laravel Mysql Ubuntu Git Menu
 

Convert CSV Data to Nested JSON in Python

Tags:

I've a requirement to convert CSV Data File to Nested JSON in Python for an application. My below current Python code is working fine for 1 Customer/Accounts document but somehow not able to create json dump for all the customers in the CSV file.

I am providing the Python code below which will give you some insight on what I am trying to achieve. Please let me know if any existing resolutions on this.

Sample Python CODE:

import pandas as pd
from itertools import groupby 
from collections import OrderedDict
import json    

df = pd.read_csv('cust.csv', dtype={
        "ClientID" : str,
        "ClientName" : str,
        "AcctID" : str,
        "AcctNbr" : str,
        "AcctTyp" : str
    })

results = []

for (ClientID, ClientName), bag in df.groupby(["ClientID", "ClientName"]):
contents_df = bag.drop(["ClientID", "ClientName"], axis=1)
subset = [OrderedDict(row) for i,row in contents_df.iterrows()]
results.append(OrderedDict([("ClientID", ClientID),("ClientName", ClientName),("subset", subset)]))

print json.dumps(results[0], indent=4)

with open('ExpectedJsonFile.json', 'w') as outfile:
outfile.write(json.dumps(results[0], indent=4))

Sample INPUT CSV:

ClientID,ClientName,AcctID,AcctNbr,AcctTyp
----------------------------------------------------------
00001,John George,812001,812001095,DDA
00001,John George,813002,813002096,SAV
00001,John George,814003,814003097,AFS
00024,Richard Polado,512987,512987085,ML
00024,Richard Polado,512983,512983086,IL
00345,John Cruze,1230,123001567,SAV
00345,John Cruze,5145,514502096,CD
00345,John Cruze,7890,7890033527,SGD

Desired Output JSON:

{  
   "clientId":00001,
   "ClientName":"John George",
   "subset":[  
      {  
         "AcctID":812001,
         "AcctNbr":"812001095",
         "AcctTyp":"DDA",
      },
      {  
         "AcctID":813002,
         "AcctNbr":"813002096",
         "AcctTyp":"SAV",
      },
      {  
         "AcctID":814003,
         "AcctNbr":"814003097",
         "AcctTyp":"AFS",
      }
   ]
},
{  
   "clientId":00024,
   "ClientName":"Richard Polado",
   "subset":[  
      {  
         "AcctID":512987,
         "AcctNbr":"512987085",
         "AcctTyp":"ML",
      },
      {  
         "AcctID":512983,
         "AcctNbr":"512983086",
         "AcctTyp":"IL",
      }
   ]
}

and these documents should continue creating for other thousands of clients.

like image 847
ShivKumar Avatar asked Jun 29 '18 06:06

ShivKumar


People also ask

Can you convert CSV to JSON?

JSON requires the data to be in a structure or a schema, which are not compatible with the CSV file structure. CSV to JSON Converter tool is designed to convert CSV files into JSON format in a very easy manner.


1 Answers

The solution groups by every 'ClientID','ClientName' pair

Your DataFrame

df = pd.DataFrame([['00001','John George','812001','812001095','DDA'],
['00001','John George','813002','813002096','SAV'],
['00001','John George','814003','814003097','AFS'],
['00024','Richard Polado','512987','512987085','ML'],
['00024','Richard Polado','512983','512983086','IL'],
['00345','John Cruze','1230','123001567','SAV'],
['00345','John Cruze','5145','514502096','CD'],
['00345','John Cruze','7890','7890033527','SGD']])

df.columns = ['ClientID','ClientName','AcctID','AcctNbr','AcctTyp'] 

and now

finalList = []
finalDict = {}
grouped = df.groupby(['ClientID', 'ClientName'])
for key, value in grouped:


    dictionary = {}

    j = grouped.get_group(key).reset_index(drop=True)
    dictionary['ClientID'] = j.at[0, 'ClientID']
    dictionary['ClientName'] = j.at[0, 'ClientName']


    dictList = []
    anotherDict = {}
    for i in j.index:

        anotherDict['AcctID'] = j.at[i, 'AcctID']
        anotherDict['AcctNbr'] = j.at[i, 'AcctNbr']
        anotherDict['AcctTyp'] = j.at[i, 'AcctTyp']

        dictList.append(anotherDict)

    dictionary['subset'] = dictList


    finalList.append(dictionary)

import json
json.dumps(finalList)

gives:

'[
   {"ClientID": "00001", 
    "ClientName": "John George", 
    "subset": 
            [{"AcctID": "814003", 
              "AcctNbr": "814003097", 
              "AcctTyp": "AFS"}, 

             {"AcctID": "814003", 
              "AcctNbr": "814003097", 
              "AcctTyp": "AFS"}, 

             {"AcctID": "814003", 
              "AcctNbr": "814003097", 
              "AcctTyp": "AFS"}]

   }, 

  {
   "ClientID": "00024", 
   "ClientName": "Richard Polado", 
   "subset": 
            [{"AcctID": "512983", 
              "AcctNbr": "512983086", 
              "AcctTyp": "IL"}, 

             {"AcctID": "512983", 
              "AcctNbr": "512983086", 
              "AcctTyp": "IL"}]
   }, 

  {
   "ClientID": "00345", 
   "ClientName": "John Cruze", 
   "subset": 
            [{"AcctID": "7890", 
              "AcctNbr": "7890033527", 
              "AcctTyp": "SGD"}, 

             {"AcctID": "7890", 
              "AcctNbr": "7890033527", 
              "AcctTyp": "SGD"}, 

             {"AcctID": "7890", 
              "AcctNbr": "7890033527", 
              "AcctTyp": "SGD"}]
   }

]'

Is this how you seem to want it?

like image 98
imperialgendarme Avatar answered Oct 11 '22 19:10

imperialgendarme