What is the simplest way to compare two columns using groupby - without apply/lambda?

df  product1 product2  person
0    apple   apples     abby
1    apple   apple      abby
...

df.groupby("person").product1.eq(df.groupby("person").product2)

Expected Output

0  False
1  True

CodePudding user response：

A bit tricky to solve without apply but you can use merge to compare ('person', 'product2) to ('person', 'product1'):

idx = (df.rename(columns={'product2': 'product'})[['person', 'product']].reset_index()
         .merge(df.rename(columns={'product1': 'product'})[['person', 'product']], 
                on=['person', 'product'])['index'].unique())

df['compare'] = df.index.isin(idx)
print(df)

# Output
  product1 product2 person  compare
0    apple   apples   abby    False
1    apple    apple   abby     True