fill a pandas column with a non-empty condition-CodePudding

I've got two panas dataframe, and I would like to generate one more with them:

import pandas as pd


df_name=pd.DataFrame({'CODE':['01','02','03','04'],
'NAME':['MICK','DAVID','JEAN','CHERRY']})

df_exp=pd.DataFrame({'TEST':['POST1','POST2','POST3','POST4','POST5','POST6','POST7', 'POST8'],
'PERS CODE':['02','','','04','','02','','03'],
'PERS2 CODE':['03','01','02','01','04','02','01','05']})

print(df_name)
print(df_exp)

df_output=pd.DataFrame()
df_output['POST']=df_exp['TEST']
df_output['AGENT'] = df_exp['PERS CODE'].map(df_name.set_index('CODE')['NAME'])

print(df_output)

but I would like a condition: if the value is empty in 'PERS CODE' it take the Value of 'PERS2 CODE'

CodePudding user response：

You can use mask to fill the empty 'PERS CODE' entries with those of 'PERS2 CODE':

df_output['AGENT'] = df_exp['PERS CODE'].mask(df_exp['PERS CODE'].eq(''), df_exp['PERS2 CODE']).map(df_name.set_index('CODE')['NAME'])