How do I combine three text column in pandas using if & else-CodePudding

I'm using python 3.7 I have a pandas data frame with three text columns,name,email & section. The sample data looks like

Name   Email              Section
abc    [email protected]      purchase
cde      -                drawing
lmn-pqr      None               -

Hyphen are there in between two words in all of the three columns. I would like to join three columns with "_" as separator and create a new column group ignoring None or -. My combined outcome will look like

Name   Email              Section   Group
abc    [email protected]      purchase  [email protected]_purchase
cde      -                drawing   cde_drawing
lmn-pqr      None               -   lmn-pqr

I'm not sure about the python code. Can you please help me?

CodePudding user response：

You can use str.cat that gets rid of null values:

df.mask(df.isin(['-', None])).apply(lambda r: r.str.cat(sep='_'), axis=1)

or, manually:

df['Group'] = df.apply(lambda r: '_'.join([x for x in r.replace('-', pd.NA).dropna()]),
                       axis=1)

output:

      Name          Email   Section                       Group
0      abc  [email protected]  purchase  [email protected]_purchase
1      cde              -   drawing                 cde_drawing
2  lmn-pqr           None         -                     lmn-pqr

CodePudding user response：

You can try replace - with None then filter it out when join

df['Group'] = df.replace({'-': None}).apply(lambda row: '_'.join(filter(None, row)), axis=1)

print(df)

      Name          Email   Section                       Group
0      abc  [email protected]  purchase  [email protected]_purchase
1      cde              -   drawing                 cde_drawing
2  lmn-pqr           None         -                     lmn-pqr

CodePudding user response：

df['Group'] = df.apply(lambda x: '-'.join([x['Name'], x['Email'], x['Section']))

x is a Series.