可以使用get_dummies和groupby.sum：

out = df[['# LOAN']].join(pd.get_dummies(df[['SECTOR', 'COUNTRY']].stack())
                            .groupby(level=0).sum())

• 注意：如果两列有可能具有相同的值，则使用.groupby(level=0).max().astype(int)。*

输出：

# LOAN  auto  food  germany  italy  japan  telecom
0  Loan 1     0     1        1      0      0        0
1  Loan 2     0     0        0      1      0        1
2  Loan 3     1     0        0      0      1        0
3  Loan 4     0     1        0      0      1        0
4  Loan 5     0     0        1      0      0        1
5  Loan 6     1     0        0      1      0        0

备选方案：
关于str.get_dummies

out = df[['# LOAN']].join(df[['SECTOR', 'COUNTRY']]
                          .agg('|'.join, axis=1)
                          .str.get_dummies()
                          )

crosstab

cols = ['SECTOR', 'COUNTRY']

out = (pd.concat(pd.crosstab(df['# LOAN'], df[c]) for c in cols)
         .groupby(level=0).sum().reset_index()
       )

如果你使用sklearn，你可以使用OneHotEncoder和ColumnTransformer：

from sklearn.preprocessing import OneHotEncoder
from sklearn.compose import ColumnTransformer

ct = ColumnTransformer(
    [('OHE', OneHotEncoder(dtype=int), ['SECTOR', 'COUNTRY'])],
    remainder='passthrough', verbose_feature_names_out=False
)
out = pd.DataFrame(ct.fit_transform(df), columns=ct.get_feature_names_out())

输出：

>>> out
  SECTOR_auto SECTOR_food SECTOR_telecom COUNTRY_germany COUNTRY_italy COUNTRY_japan  # LOAN
0           0           1              0               1             0             0  Loan 1
1           0           0              1               0             1             0  Loan 2
2           1           0              0               0             0             1  Loan 3
3           0           1              0               0             0             1  Loan 4
4           0           0              1               1             0             0  Loan 5
5           1           0              0               0             1             0  Loan 6

你要找的是独热编码。有一个关于如何从pd.DataFrame（）中获取的很棒的线程：How can I one hot encode in Python?
控制论的回答相当彻底
编辑：mozway在这个线程上的回答是完全正确的--getdummy的是pandas fxn的one-hot编码，我相信他们在链接的线程中使用了相同的编码