CODEWITHSUNDEEP
pythonpandasdataframesplitmultiple-columns
dagg3r
dagg3r

Reputation: 341

Pandas add new columns based on splitting another column

I have a pandas dataframe like the following:

A              B
US,65,AMAZON   2016
US,65,EBAY     2016

My goal is to get to look like this:

A              B      country    code    com
US.65.AMAZON   2016   US         65      AMAZON
US.65.AMAZON   2016   US         65      EBAY

I know this question has been asked before here and here but none of them works for me. I have tried: 

df['country','code','com'] = df.Field.str.split('.')

and

df2 = pd.DataFrame(df.Field.str.split('.').tolist(),columns = ['country','code','com','A','B'])

Am I missing something? Any help is much appreciated.

Upvotes: 7

Views: 13515

Answers (3)

user10451754
user10451754

Reputation: 11

This will not give the output as expected it will only give the df['A'] first value which is 'U'

This is okay to create column based on provided data df1=pd.DataFrame([x.split(',') for x in df['A'].tolist()],columns= ['country','code','com'])

instead of for lambda also can be use

Upvotes: 1

Nithin Narla
Nithin Narla

Reputation: 21

For getting the new columns I would prefer doing it as following:

df['Country'] = df['A'].apply(lambda x: x[0])
df['Code'] = df['A'].apply(lambda x: x[1])
df['Com'] = df['A'].apply(lambda x: x[2])

As for the replacement of , with a . you can use the following:

df['A'] = df['A'].str.replace(',','.')

Upvotes: 1

jezrael
jezrael

Reputation: 862591

You can use split with parameter expand=True and add one [] to left side:

df[['country','code','com']] = df.A.str.split(',', expand=True)

Then replace , to .:

df.A = df.A.str.replace(',','.')

print (df)
              A     B country code     com
0  US.65.AMAZON  2016      US   65  AMAZON
1    US.65.EBAY  2016      US   65    EBAY

Another solution with DataFrame constructor if there are no NaN values:

df[['country','code','com']] = pd.DataFrame([ x.split(',') for x in df['A'].tolist() ])
df.A = df.A.str.replace(',','.')
print (df)
              A     B country code     com
0  US.65.AMAZON  2016      US   65  AMAZON
1    US.65.EBAY  2016      US   65    EBAY

Also you can use column names in constructor, but then concat is necessary:

df1=pd.DataFrame([x.split(',') for x in df['A'].tolist()],columns= ['country','code','com'])
df.A = df.A.str.replace(',','.')
df = pd.concat([df, df1], axis=1)
print (df)
              A     B country code     com
0  US.65.AMAZON  2016      US   65  AMAZON
1    US.65.EBAY  2016      US   65    EBAY

Upvotes: 9

Related Questions