如何将 Dataframe Multiindex 中的值与相同索引的系列进行比较

Question

我有一个名为 chunk 的数据框，如下所示：

                                         Expiration   DataDate durations  
UnderlyingSymbol Delta   OptionSymbol                                         
A                 0.9991 32500           2020-05-15 2020-05-01   14 days   
                         35000           2020-05-15 2020-05-01   14 days   
                         37500           2020-05-15 2020-05-01   15 days   
                         40000           2020-05-15 2020-05-01   16 days   
                         42500           2020-05-15 2020-05-01   14 days   
                         45000           2020-05-15 2020-05-01   13 days   
                 -0.9152 15000           2020-11-20 2020-05-01  203 days      
AAL              -0.9142 20000           2020-06-05 2020-05-01   35 days

我有一个名为 durations_star 的系列，其 multindex 与数据框中的前两个级别匹配：

UnderlyingSymbol  Delta  
A                  0.9991    14 days
                  -0.9152   203 days
AAL               -0.9142    35 days
Name: durations, dtype: timedelta64[ns]

有没有办法 select 数据框中的行，对于每个（UnderlyingSymbol，Delta），都等于（一旦转换为整数）系列中的相应值？例如

                                         Expiration   DataDate durations  \
UnderlyingSymbol Delta   OptionSymbol                                         
A                 0.9991 32500           2020-05-15 2020-05-01   14 days   
                         35000           2020-05-15 2020-05-01   14 days   
                         42500           2020-05-15 2020-05-01   14 days   
                 -0.9152 15000           2020-11-20 2020-05-01  203 days      
AAL              -0.9142 20000           2020-06-05 2020-05-01   35 days

有没有更好的方法：

chunk.reset_index().groupby(['UnderlyingSymbol', 'Delta']).apply(lambda t : t[ t.durations == durations_star[t.name[0], t.name[1] ] ] )

Answer 1

您可以使用 join:

>>> df.join(sr.rename('durations2')) \
      .query('durations == durations2') \
      .drop(columns='durations2')

                                       Expiration    DataDate durations
UnderlyingSymbol Delta   OptionSymbol                                  
A                -0.9152 15000         2020-11-20  2020-05-01  203 days
                  0.9991 32500         2020-05-15  2020-05-01   14 days
                         35000         2020-05-15  2020-05-01   14 days
                         42500         2020-05-15  2020-05-01   14 days
AAL              -0.9142 20000         2020-06-05  2020-05-01   35 days

设置：

# Your DataFrame
>>> df
                                       Expiration    DataDate durations
UnderlyingSymbol Delta   OptionSymbol                                  
A                 0.9991 32500         2020-05-15  2020-05-01   14 days
                         35000         2020-05-15  2020-05-01   14 days
                         37500         2020-05-15  2020-05-01   15 days
                         40000         2020-05-15  2020-05-01   16 days
                         42500         2020-05-15  2020-05-01   14 days
                         45000         2020-05-15  2020-05-01   13 days
                 -0.9152 15000         2020-11-20  2020-05-01  203 days
AAL              -0.9142 20000         2020-06-05  2020-05-01   35 days


# Your Series
>>> sr
UnderlyingSymbol  Delta  
A                  0.9991     14 days
                  -0.9152    203 days
AAL               -0.9142     35 days
Name: durations, dtype: object

如何将 Dataframe Multiindex 中的值与相同索引的系列进行比较

How to compare values in Dataframe Multindex with equally indexed series

pandas

select

multi-index