python - pandas 中多列分类值的映射

Question

假设我有一个包含三列分类数据的数据框，我想将三个分类列转换为单个值并映射到原始数据框。我知道这可以通过带有 this 的单列来实现。 ，但是多列怎么样？

示例:从此

>>>df = pd.DataFrame({'id':['0', '1', '2', '3','4'],
...                   'x':['tall', 'short', 'tall', 'short', 'tall'],
...                   'y':['fat', 'thin', 'thin', 'fat', 'fat'],
...                   'z':['male', 'female', 'female', 'male', 'male']},
...                   dtype='category')

>>>df
  id      x     y       z
0  0   tall   fat    male
1  1  short  thin  female
2  2   tall  thin  female
3  3  short   fat    male
4  4   tall   fat    male

通过映射 x、y 和 z 列来实现此目的

>>>df
  id      x     y       z  map
0  0   tall   fat    male    0
1  1  short  thin  female    1
2  2   tall  thin  female    2
3  3  short   fat    male    3
4  4   tall   fat    male    0

Answer 1

这是groupby().ngroup():

df['map'] = df.groupby(['x','y','z'], sort=False).ngroup()

或者，如果您的数据是字符串类型，您可以连接列(可能带有一些特殊字符)，并使用单列方法:

# add('&') may not be needed
df['map'] = pd.factorize(df[['x','y','z']].add('&').sum(1))[0]

输出:

   id      x     y       z  map
0   0   tall   fat    male    0
1   1  short  thin  female    1
2   2   tall  thin  female    2
3   3  short   fat    male    3
4   4   tall   fat    male    0