pandas merge two columns with customized text

Question

I am having following dataframe

df1 = pd.DataFrame({'Name': ['A0', 'A1', 'A2', 'A3', 'A4'],
                    'Buy': [True, True, False, False, False],
                    'Sell': [False, False, True, False, True]
                   },
                   index=[0, 1, 2, 3, 4])
df1

    Name    Buy Sell
0   A0  True    False
1   A1  True    False
2   A2  False   True
3   A3  False   False
4   A4  False   True

I want to merge Buy and Sell columns on a condition that if "Buy" is having True value then "Buyer" if "Sell" has True value then "Seller" and if both "Buy" and "Sell" has False value then it should have "NA"

sample required output

    Name    Type 
0   A0      Buyer
1   A1      Buyer
2   A2      Seller
3   A3      NA
4   A4      Seller

Answer 1

np.select

a = np.select([df1.Buy, df1.Sell], ['Buyer', 'Seller'], 'NA')
pd.DataFrame({'Name': df1.Name, 'Type': a})

  Name    Type
0   A0   Buyer
1   A1   Buyer
2   A2  Seller
3   A3      NA
4   A4  Seller

---

df1.assign(Type=np.select([df1.Buy, df1.Sell], ['Buyer', 'Seller'], 'NA'))

  Name    Buy   Sell    Type
0   A0   True  False   Buyer
1   A1   True  False   Buyer
2   A2  False   True  Seller
3   A3  False  False      NA
4   A4  False   True  Seller

Answer 2

You can do:

s = df[['Buy','Sell']]
df['Type'] = ([email protected]).add('er').replace('er', np.nan)

# or 
# df['Type'] = np.where(s.any(1), s.idxmax(1).add('er'), np.nan)

Output:

  Name    Buy   Sell    Type
0   A0   True  False   Buyer
1   A1   True  False   Buyer
2   A2  False   True  Seller
3   A3  False  False     NaN
4   A4  False   True  Seller