python支持获取特定值的多级索引

我每小时有一个分组值的数据帧

2020.05.05 17:00:00  cpu_usage_specint        4.354603
                     phys_iops             3075.779680
                     total_memory         13843.480000
2020.05.05 18:00:00  cpu_usage_specint        3.670237
                     phys_iops             3403.007340
                                              ...     
2020.05.11 22:00:00  phys_iops            30459.577340
                     total_memory         28851.025000
2020.05.11 23:00:00  cpu_usage_specint        1.903852
                     phys_iops             8161.811320
                     total_memory         28827.575000

the index is a MultiIndex([('2020.05.05 17:00:00', 'cpu_usage_specint'),
            ('2020.05.05 17:00:00',         'phys_iops'),
            ('2020.05.05 17:00:00',      'total_memory'),
            ('2020.05.05 18:00:00', 'cpu_usage_specint'),
            ('2020.05.05 18:00:00',         'phys_iops'),
            ('2020.05.05 18:00:00',      'total_memory'),
            ('2020.05.05 19:00:00', 'cpu_usage_specint'),
            ('2020.05.05 19:00:00',         'phys_iops'),
            ('2020.05.05 19:00:00',      'total_memory'),
            ('2020.05.05 20:00:00', 'cpu_usage_specint'),
            ...
            ('2020.05.11 20:00:00',      'total_memory'),
            ('2020.05.11 21:00:00', 'cpu_usage_specint'),
            ('2020.05.11 21:00:00',         'phys_iops'),
            ('2020.05.11 21:00:00',      'total_memory'),
            ('2020.05.11 22:00:00', 'cpu_usage_specint'),
            ('2020.05.11 22:00:00',         'phys_iops'),
            ('2020.05.11 22:00:00',      'total_memory'),
            ('2020.05.11 23:00:00', 'cpu_usage_specint'),
            ('2020.05.11 23:00:00',         'phys_iops'),
            ('2020.05.11 23:00:00',      'total_memory')],

例如，我想要每个小时的每个度量的具体值

for date in metrics_pivot_df.index.get_level_values('date').unique():
    print('this is the date',date)

这给了我日期，但我也想要总内存的值，这样我就可以做一些测试，比如>=等等，但是我不知道如何获得该值？我曾尝试引用“mertic”列，但当我尝试执行查询时，返回true/false

---

非常感谢任何帮助

我看不见树木

for date in metrics_pivot_df.index.get_level_values('date').unique():
    print('this is a specific date',date, metrics_pivot_df.loc[date, 'cpu_usage_specint'])

解决了