python - Pandas 按索引分组并计算总和

Question

我有以下系列:

s = pd.Series({
    (0, 'InData'): 2944.51697, (0, 'InInterests'): 3703.208935,
    (1, 'InData'): 6207.58004, (1, 'InInterests'): 3505.068411,
    (2, 'InData'): 8133.00273, (2, 'InInterests'): 3735.028306,
    (3, 'InData'): 9426.70471, (3, 'InInterests'): 2665.989291,
    (4, 'InData'): 2604.1578, (4, 'InInterests'): 3103.310729,
    (5, 'InData'): 6784.82916, (5, 'InInterests'): 3293.815375,
    (6, 'InData'): 6823.6174, (6, 'InInterests'): 4121.83398,
    (7, 'InData'): 4072.70277, (7, 'InInterests'): 3033.609368,
    (8, 'InData'): 4614.60824, (8, 'InInterests'): 2955.216811,
    (9, 'InData'): 6986.50075, (9, 'InInterests'): 2986.820394
}, name='KilobytesRaw').rename_axis(index=['Node', 'Type'])

:

Node  Type       
0     InData         2944.516970
      InInterests    3703.208935
1     InData         6207.580040
      InInterests    3505.068411
2     InData         8133.002730
      InInterests    3735.028306
3     InData         9426.704710
      InInterests    2665.989291
4     InData         2604.157800
      InInterests    3103.310729
5     InData         6784.829160
      InInterests    3293.815375
6     InData         6823.617400
      InInterests    4121.833980
7     InData         4072.702770
      InInterests    3033.609368
8     InData         4614.608240
      InInterests    2955.216811
9     InData         6986.500750
      InInterests    2986.820394
Name: KilobytesRaw, dtype: float64

我想计算 InData 和 InInterests 的总和，但在 Pandas 中找不到这种情况 indexing page ，也不在 Google 上。

所以我希望我的结果看起来像这样:

Node
0     6647.725905
1     9712.648451
...
9     9973.321144

Answer 1

Groupby sum支持传递级别编号而不是列名称。这允许在一个级别而不是一个列上进行求和:

s.groupby(level=0).sum()

既然索引级别是命名的，我们也可以使用索引名称代替级别编号:

s.groupby(level='Node').sum()

任一选项都会产生:

Node
0     6647.725905
1     9712.648451
2    11868.031036
3    12092.694001
4     5707.468529
5    10078.644535
6    10945.451380
7     7106.312138
8     7569.825051
9     9973.321144
Name: KilobytesRaw, dtype: float64