python - Pandas 多索引: print all of first index if in second index

Question

我想查询数据帧并输出第一个索引中的所有项目(如果它包含在第二个索引中)。 描述我想要实现的目标的简化版本是:

data = {'colour': ['red','purple','green','purple','blue','red'], 'item': ['hat','scarf','belt','belt','hat','scarf'], 'material': ['felt','wool','leather','wool','plastic','wool']}
df = pd.DataFrame(data=data)
grpd_df = df.groupby(df['item']).apply(lambda df: df.reset_index(drop=True))
grpd_df

         colour  item material
item

belt   0 green   belt  leather 
       1 purple  belt  wool 

hat    0 red     hat   felt 
       1 blue    hat   plastic 

scarf  0 purple  scarf wool 
       1 red     scarf wool 

我想获取项目中包含红色项目的所有行:

hat    0 red     hat   felt 
       1 blue    hat   plastic 

scarf  0 purple  scarf wool 
       1 red     scarf wool 

Answer 1

使用groupby通过将 color 列与 eq 与 2 系列进行比较与 any每组至少有一个 True:

df = grpd_df[grpd_df['colour'].eq('red').groupby(level=0).transform('any')]
print (df)
         colour   item material
item                           
hat   0     red    hat     felt
      1    blue    hat  plastic
scarf 0  purple  scarf     wool
      1     red  scarf     wool

详细信息:

print (grpd_df['colour'].eq('red').groupby(level=0).transform('any'))
item    
belt   0    False
       1    False
hat    0     True
       1     True
scarf  0     True
       1     True
Name: colour, dtype: bool

较慢的替代方案 filter :

df = grpd_df.groupby(level=0).filter(lambda x: x['colour'].eq('red').any())

如果想使用原始DataFrame:

df = df[df['colour'].eq('red').groupby(df['item']).transform('any')]
print (df)
   colour   item material
0     red    hat     felt
1  purple  scarf     wool
4    blue    hat  plastic
5     red  scarf     wool

编辑:

如果想使用MultiIndex:

data = {'colour': ['red','purple','green','purple','blue','red'], 'item': ['hat','scarf','belt','belt','hat','scarf'], 'material': ['felt','wool','leather','wool','plastic','wool']}
df = pd.DataFrame(data=data).set_index(['colour','item'])

print (df)    
             material
colour item          
red    hat       felt
purple scarf     wool
green  belt   leather
purple belt      wool
blue   hat    plastic
red    scarf     wool

df = df[pd.Series(df.index.get_level_values('colour') == 'red', index=df.index).groupby(level=1).transform('any')]

第二个过滤器解决方案:

df = df.groupby(level=1).filter(lambda x: (x.index.get_level_values('colour') == 'red').any())

---

print (df)

             material
colour item          
red    hat       felt
purple scarf     wool
blue   hat    plastic
red    scarf     wool