python - 以 MultiIndex 值为条件沿 Pandas Column 求和？

Question

我有以下 Pandas DataFrame df:

                                 Value
time                 Position
1493791210867023000  0.0           21156.0
                     1.0         1230225.0
                     2.0         1628088.0
                     3.0         2582359.0
                     4.0         3388164.0
1493791210880251000  0.0           21156.0
                     1.0         1230225.0
                     2.0         1628088.0
                     3.0         2582359.0
                     4.0         3388164.0
1493791210888418000  0.0           21156.0
                     1.0         1230225.0
...                  ...         ...

我怎样才能有效地沿着索引“位置”求和？我要实现的确切求和公式是:

                                 Value        Result
time                 Position
1493791210867023000  0.0           21156.0    Sum from 0.0 to 0.0
                     1.0         1230225.0    Sum from 0.0 to 1.0
                     2.0         1628088.0    Sum from 0.0 to 2.0
                     3.0         2582359.0    Sum from 0.0 to 3.0
                     4.0         3388164.0    Sum from 0.0 to 4.0
1493791210880251000  0.0           21156.0    Sum from 0.0 to 0.0
                     1.0         1230225.0    Sum from 0.0 to 1.0
                     2.0         1628088.0    Sum from 0.0 to 2.0
                     3.0         2582359.0    Sum from 0.0 to 3.0
...                  ...         ...          ...

我当前的解决方案花费的时间太长(IndexSlice 非常慢)而且我不太确定如何将总和结果有效地排序到(新创建的)“结果”列中？

import pandas as pd
import numpy as np
idx = pd.IndexSlice
res = {}
for i in range(5):
    res[i] = df.loc[idx[:, :i]].groupby(level="time").sum()
df["Result"] = 0 #fill Result now with res[i] depending on position

Answer 1

尝试在 groupby 中使用 cumsum

df.assign(Result=df.groupby(level='time').Value.cumsum())
# suggested by @ScottBoston for pandas 0.20.1+
# df.assign(Result=df.groupby('time').Value.cumsum())

                                  Value     Result
time                Position                      
1493791210867023000 0.0         21156.0    21156.0
                    1.0       1230225.0  1251381.0
                    2.0       1628088.0  2879469.0
                    3.0       2582359.0  5461828.0
                    4.0       3388164.0  8849992.0
1493791210880251000 0.0         21156.0    21156.0
                    1.0       1230225.0  1251381.0
                    2.0       1628088.0  2879469.0
                    3.0       2582359.0  5461828.0
                    4.0       3388164.0  8849992.0
1493791210888418000 0.0         21156.0    21156.0
                    1.0       1230225.0  1251381.0