How to convert json into a pandas dataframe?

Question

I'm trying to covert an api response from json to a dataframe in pandas. the problem I am having is that de data is nested in the json format and I am not getting the right columns in my dataframe.

The data is collect from a api with the following format:

{'tickets': [{'url': 'https...',
   'id': 1,
   'external_id': None,
   'via': {'channel': 'web',
    'source': {'from': {}, 'to': {}, 'rel': None}},
   'created_at': '2020-05-01T04:16:33Z',
   'updated_at': '2020-05-23T03:02:49Z',
   'type': 'incident',
   'subject': 'Subject',
   'raw_subject': 'Raw subject',
   'description': 'Hi, this is the description',
   'priority': 'normal',
   'status': 'closed',
   'recipient': None,
   'requester_id': 409467360874,
   'submitter_id': 409126461453,
   'assignee_id': 409126461453,
   'organization_id': None,
   'group_id': 360009916453,
   'collaborator_ids': [],
   'follower_ids': [],
   'email_cc_ids': [],
   'forum_topic_id': None,
   'problem_id': None,
   'has_incidents': False,
   'is_public': True,
   'due_at': None,
   'tags': ['tag_1',
    'tag_2',
    'tag_3',
    'tag_4'],
   'custom_fields': [{'id': 360042034433, 'value': 'value of the first custom field'},
    {'id': 360041487874, 'value': 'value of the second custom field'},
    {'id': 360041489414, 'value': 'value of the third custom field'},
    {'id': 360040980053, 'value': 'correo_electrónico'},
    {'id': 360040980373, 'value': 'suscribe_newsletter'},
    {'id': 360042046173, 'value': None},
    {'id': 360041028574, 'value': 'product'},
    {'id': 360042103034, 'value': None}],
   'satisfaction_rating': {'score': 'unoffered'},
   'sharing_agreement_ids': [],
   'comment_count': 2,
   'fields': [{'id': 360042034433, 'value': 'value of the first custom field'},
    {'id': 360041487874, 'value': 'value of the second custom field'},
    {'id': 360041489414, 'value': 'value of the third custom field'},
    {'id': 360040980053, 'value': 'correo_electrónico'},
    {'id': 360040980373, 'value': 'suscribe_newsletter'},
    {'id': 360042046173, 'value': None},
    {'id': 360041028574, 'value': 'product'},
    {'id': 360042103034, 'value': None}],
   'followup_ids': [],
   'ticket_form_id': 360003608013,
   'deleted_ticket_form_id': 360003608013,
   'brand_id': 360004571673,
   'satisfaction_probability': None,
   'allow_channelback': False,
   'allow_attachments': True},

What I already tried is the following: I have converted the JSON format into a dict as following:

x = response.json()
df = pd.DataFrame(x['tickets'])

But I'm struggling with the output. I don't know how to get a correct, ordered, normalized dataframe.

(I'm new in this:) )

Answer 1

Let's supose you get your request data by this code r = requests.get(url, auth)

Your data ins't clear yet, so let's get a dataframe of it data = pd.read_json(json.dumps(r.json, ensure_ascii = False))

But, probably you will get a dataframe with one single row.

When I faced a problem like this, I wrote this function to get the full data:

listParam = []

def listDict(entry):
    if type(entry) is dict:
        listParam.append(entry)
    elif type(entry) is list:
        for ent in entry:
            listDict(ent)

Because your data looks like a dict because of {'tickets': ...} you will need to get the information like that:

listDict(data.iloc[0][0])

And then,

pd.DataFrame(listParam)

I can't show the results because you didn't post the complete data nor told where I can find the data to test, but this will probably work.

Answer 2

You have to convert the json to dictionary first and then convert the dictionary value for key 'tickets' into dataframe.

file = open('file.json').read()
ticketDictionary = json.loads(file)
df = pd.DataFrame(ticketDictionary['tickets'])

'file.json' contains your data here.

df now contains your dataFrame in this format.

For the lists within the response you can have separate dataframes if required:

for field in df['fields']:
        df = pd.DataFrame(field)

It will give you this for lengths:

            id                             value
0  360042034433   value of the first custom field
1  360041487874  value of the second custom field
2  360041489414   value of the third custom field
3  360040980053                correo_electrónico
4  360040980373               suscribe_newsletter
5  360042046173                              None
6  360041028574                           product
7  360042103034                              None

This can be one way to structure as you haven't mentioned the exact expected format.