20
votes

I have a Json file as follows. It's a list of dicts.

[{"city": "ab", "trips": 4, "date": "2014-01-25", "value": 4.7, "price": 1.1, "request_date": "2014-06-17", "medium": "iPhone", "%price": 15.4, "type": true, "Weekly_pct": 46.2, "avg_dist": 3.67, "avg_price": 5.0}, {"city": "bc", "trips": 0, "date": "2014-01-29", "value": 5.0, "price": 1.0, "request_date": "2014-05-05", "medium": "Android", "%price": 0.0, "type": false, "weekly_pct": 50.0, "avg_dist": 8.26, "avg_price": 5.0}.....]

When I read this using this:

data=pd.read_json('dataset.json')

I get the following error:

ValueError: Expected object or value

I tried this too:

from ast import literal_eval

with open('dataset.json') as f:
    data = literal_eval(f.read())

df = pd.DataFrame(data)

It gives the following error:

ValueError: malformed string

Edit:

Even Json.loads doesn't work. Tried this:

import json
data=json.loads('dataset.json')

ValueError: No JSON object could be decoded

The Json file is 13.5MB but it seems to have huge amounts of data.

6
Does json.loads work? (import json first...) - IanS
No. I tried that it gave me this error. Check Edit. - Baktaawar
I think you might benefit from this answer: stackoverflow.com/a/20644150/5276797 - IanS

6 Answers

17
votes

I think you can use modul json for reading file.json and then DataFrame constructor:

import pandas as pd
import json

with open('file.json') as f:
   data = json.load(f)
print data
[{u'city': u'ab', u'medium': u'iPhone', u'request_date': u'2014-06-17', u'price': 1.1, u'Weekly_pct': 46.2, u'value': 4.7, u'%price': 15.4, u'avg_price': 5.0, u'date': u'2014-01-25', u'avg_dist': 3.67, u'type': True, u'trips': 4}, {u'city': u'bc', u'medium': u'Android', u'request_date': u'2014-05-05', u'price': 1.0, u'weekly_pct': 50.0, u'value': 5.0, u'%price': 0.0, u'avg_price': 5.0, u'date': u'2014-01-29', u'avg_dist': 8.26, u'type': False, u'trips': 0}]

print pd.DataFrame(data)

   %price  Weekly_pct  avg_dist  avg_price city        date   medium  price  \
0    15.4        46.2      3.67        5.0   ab  2014-01-25   iPhone    1.1   
1     0.0         NaN      8.26        5.0   bc  2014-01-29  Android    1.0   

  request_date  trips   type  value  weekly_pct  
0   2014-06-17      4   True    4.7         NaN  
1   2014-05-05      0  False    5.0        50.0  
8
votes

You need to indicate to Pandas that "records" formatting (where the JSON appears like a list of dictionaries) is used in datasets.json.

res = pd.read_json('input/dataset.json', orient='records')

print(res.iloc[:, :5])
   %price  Weekly_pct  avg_dist  avg_price city
0    15.4        46.2      3.67          5   ab
1     0.0         NaN      8.26          5   bc
8
votes

I had the same error. Turns out it couldn't find the file. I modified the path and pd.read_json worked fine. As for json.loads, this might be helpful.

4
votes

The following worked for me when pd.read_json failed: open file, load with normal json.load, then load into a pandas dataframe.

    import pandas as pd
    import json

    openfile=open('file.json')
    jsondata=json.load(openfile)
    df=pd.DataFrame(jsondata)

    openfile.close()
    print(df)
0
votes

For me it was a problem with the path. The path I had to use depended on the directory from where I run the python file. Maybe try to 'cd' into the directory of your python file and then data=pd.read_json('dataset.json') should work.

0
votes

I had to add the parameter lines=True to make it work, e.g:

pd.read_json("dataset.json", lines=True)

Alternatively you could do it like this:

import json
import pandas as pd

with open("dataset.json") as f:
  df = pd.DataFrame([json.loads(l) for l in f.readlines()])
print(df)  # Shows data frame as expected