CODEWITHSUNDEEP
pythonpandasdataframe
Daniel Isaac
Daniel Isaac

Reputation: 745

Pandas DataFrame: Unusual Behaviour with json.dumps ( Extra double quotes )

I have this Pandas DataFrame with two columns label and time

>>> df = pd.DataFrame([{'a':{'tier':'one','app':'frontend'},'time':100}])
>>> df
                                        a  time
0  {u'tier': u'one', u'app': u'frontend'}   100

Column label stores a dict.

When I print the dataframe, I get the expected row values

>>> print(df.to_csv(index=False,header=False,sep='|'))
{'tier': 'one', 'app': 'frontend'}|100

I wanted to convert these row JSON values to a string and hence I did

>>> df['a'] = df['a'].apply(lambda x: json.dumps(x))
>>> df
                                    a  time
0  {"tier": "one", "app": "frontend"}   100

But with df.to_csv(), I get this issue where I get two times double quotes

>>> print(df.to_csv(index=False,header=False,sep='|'))
"{""tier"": ""one"", ""app"": ""frontend""}"|100

When the expected output should be 

{"tier": "one", "app": "frontend"}|100

This behaviour seems very unusual. Am I going wrong somewhere here ?

Upvotes: 3

Views: 888

Answers (1)

Bharat Gera
Bharat Gera

Reputation: 820

Use quoting=csv.QUOTE_NONE in to_csv function, as:

import csv
print(df.to_csv(index=False,header=False,sep='|',quoting=csv.QUOTE_NONE))

For more details, read the docs.

Upvotes: 3

Related Questions