带有多个索引的Pandas数据透视表

您需要reindex按unique列的值Loc_Time与sort_values按列的值Time，因为pivot默认情况下排序：

df1 = df.pivot('Group','Loc_Time','Value')
df1 = df1.reindex(columns=df.sort_values('Time')['Loc_Time'].unique())
print (df1)
Loc_Time  loc1_Week1  loc2_Week1  loc3_Week1  loc1_Week2  loc2_Week2  \
Group                                                                  
prod1             10          12          14          11          13   
prod2             20          22          24          21          23   

Loc_Time  loc3_Week2  
Group                 
prod1             15  
prod2             25  

详细说明：

print (df.sort_values('Time')['Loc_Time'].unique())

['loc1_Week1' 'loc2_Week1' 'loc3_Week1' 'loc1_Week2' 'loc2_Week2'
 'loc3_Week2']