[FIXED] error accessing the columns of a json output using pandas

Issue

I have a json output from an API like this

 {'trainingInput': 
  {'pythonModule': 'marketprice_aitp.predict', 
   'region': 'us-central1', 
   'runtimeVersion': '2.4', 
   , 
   'pythonVersion': '3.7', 
   'serviceAccount': 'projects/hh'}, 
  
  'createTime': '2021-11-08T06:49:34Z', 
  'startTime': '2021-11-08T07:07:33Z', 
  'endTime': '2021-11-08T07:07:33Z', 
  'state': 'SUCCEEDED', 
  'trainingOutput': {'consumedMLUnits': 0.09}, 
  'etag': '+YVkodd5cKA='}

I have a pandas dataframe which stores this output. And I am trying to access specific columns like belkow

df['region'] = df['trainingInput']["region"]
df['serviceaccount'] = df['trainingInput']["serviceAccount"]

I tried double quotes as well as

df['region'] = job[''''trainingInput''''][''''region''''']
df['serviceaccount'] = job[''''trainingInput''''][''''serviceAccount''''']

But I am getting key error. Please let me know your thoughts.

Solution

I actually manipulated my dataframe to convert that to correct dict format.

k_1  = df['trainingInput'].values
for i in k_1:
    m= json.dumps(i)
    k = json.loads(m)
    df['serviceaccount'] =k['serviceAccount']

And it worked.

Answered By - Maa