Conversion from nested json to csv with pandas

Question

I am trying to convert a nested json into a csv file, but I am struggling with the logic needed for the structure of my file: it's a json with 2 objects and I would like to convert into csv only one of them, which is a list with nesting.

I've found very helpful "flattening" json info in this blog post. I have been basically adapting it to my problem, but it is still not working for me.

My json file looks like this:

{
  "tickets":[
    {
      "Name": "Liam",
      "Location": {
        "City": "Los Angeles",
        "State": "CA"
      },
      "hobbies": [
        "Piano",
        "Sports"
      ],
      "year" : 1985,
      "teamId" : "ATL",
      "playerId" : "barkele01",
      "salary" : 870000
    },
    {
      "Name": "John",
      "Location": {
        "City": "Los Angeles",
        "State": "CA"
      },
      "hobbies": [
        "Music",
        "Running"
      ],
      "year" : 1985,
      "teamId" : "ATL",
      "playerId" : "bedrost01",
      "salary" : 550000
    }
  ],
  "count": 2
}

my code, so far, looks like this:

import json
from pandas.io.json import json_normalize
import argparse


def flatten_json(y):
    out = {}

    def flatten(x, name=''):
        if type(x) is dict:
            for a in x:
                flatten(x[a], name + a + '_')
        elif type(x) is list:
            i = 0
            for a in x:
                flatten(a, name + str(i) + '_')
                i += 1
        else:
            out[name[:-1]] = x
    flatten(y)
    return out


if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='Converting json files into csv for Tableau processing')
    parser.add_argument(
        "-j", "--json", dest="json_file", help="PATH/TO/json file to convert", metavar="FILE", required=True)

    args = parser.parse_args()

    with open(args.json_file, "r") as inputFile:  # open json file
        json_data = json.loads(inputFile.read())  # load json content
    flat_json = flatten_json(json_data)
    # normalizing flat json
    final_data = json_normalize(flat_json)

    with open(args.json_file.replace(".json", ".csv"), "w") as outputFile:  # open csv file

        # saving DataFrame to csv
        final_data.to_csv(outputFile, encoding='utf8', index=False)

What I would like to obtain is 1 line per ticket in the csv, with headings:

Name,Location_City,Location_State,Hobbies_0,Hobbies_1,Year,TeamId,PlayerId,Salary.

I would really appreciate anything that can do the click! Thank you!

score 6 · Answer 1 · answered Jul 16 '19 at 19:19

I actually wrote a package called cherrypicker recently to deal with this exact sort of thing since I had to do it so often!

I think the following code would give you exactly what you're after:

from cherrypicker import CherryPicker
import json
import pandas as pd

with open('file.json') as file:
    data = json.load(file)

picker = CherryPicker(data)
flat = picker['tickets'].flatten().get()
df = pd.DataFrame(flat)
print(df)

This gave me the output:

  Location_City Location_State  Name hobbies_0 hobbies_1   playerId  salary teamId  year
0   Los Angeles             CA  Liam     Piano    Sports  barkele01  870000    ATL  1985
1   Los Angeles             CA  John     Music   Running  bedrost01  550000    ATL  1985

You can install the package with:

pip install cherrypicker

...and there's more docs and guidance at https://cherrypicker.readthedocs.io.

I was researching the subject and came to your answer. Is there a way to reverse the process? from flattened csv to json? Thanks — user10033434, Feb 15 '23 at 15:25
Well JSON is just a way of formatting objects of various structures so it's really just a case of taking the parsed CSV and restructuring the contents of each row into the structure you want with lists, dicts etc. Then use the `json` library to encode it as JSON. — big-o, Feb 17 '23 at 02:39

score 2 · Accepted Answer · answered Jul 04 '19 at 09:46

An you already have a function to flatten a Json object, you have just to flatten the tickets:

...
with open(args.json_file, "r") as inputFile:  # open json file
    json_data = json.loads(inputFile.read())  # load json content
final_data = pd.DataFrame([flatten_json(elt) for elt in json_data['tickets']])
...

With your sample data, final_data is as expected:

  Location_City Location_State  Name hobbies_0 hobbies_1   playerId  salary teamId  year
0   Los Angeles             CA  Liam     Piano    Sports  barkele01  870000    ATL  1985
1   Los Angeles             CA  John     Music   Running  bedrost01  550000    ATL  1985

Sachin Prabhu · Answer 3 · 2019-07-04T09:19:34.260

1

There may be a simpler solution for this. But this should work!

import json
import pandas as pd

with open('file.json') as file:
    data = json.load(file)

df = pd.DataFrame(data['tickets'])

for i,item in enumerate(df['Location']):
    df['location_city'] = dict(df['Location'])[i]['City']
    df['location_state'] = dict(df['Location'])[i]['State']

for i,item in enumerate(df['hobbies']):
    df['hobbies_{}'.format(i)] = dict(df['hobbies'])[i]

df = df.drop({'Location','hobbies'}, axis=1)

print(df)

edited Jul 04 '19 at 09:19

answered Jul 04 '19 at 09:13

Sachin Prabhu

152
2
11

1

Thanks - I needed this exact tool for extracting nested documents from a mongo datastore. Perfect. – monachus Sep 01 '20 at 15:42

Conversion from nested json to csv with pandas

3 Answers3

Linked