熊猫：MultiIndex上的复杂查询

Question

I have the following DataFrame

data = DataFrame({'k1':['one'] * 3 + ['two'] *4,
              'k2':[1, 1, 2, 3, 3, 4, 4],
              'date': [
                DT.datetime(2013,1,1,14,0),
                DT.datetime(2013,1,1,14,5),
                DT.datetime(2013,1,2,15,0),
                DT.datetime(2013,1,2,10,0),
                DT.datetime(2013,1,3,19,0),                                      
                DT.datetime(2013,1,3,18,0),
                DT.datetime(2013,1,4,16,0)]
                  })
data.set_index(['date','k1', 'k2'], inplace=True)

Does anyone know how I can query this MultiIndex to retrieve all entries which are before 2013-01-02 and have k1 = 'one' and k2 = 1?

Any help is appreciated

Thanks

Answer 1

You can use data.index.get_level_values() to filter multiIndex

>> data.iloc[data.index.get_level_values('k1') == 'one']

to filter the index based on k1 alone. To filter based on k1 and k2 use can use

>> data.iloc[(data.index.get_level_values('k1') == 'one') & (data.index.get_level_values('k2') == 1)]