Sum dataframe列前一行和两列的乘积

Question

我有一个包含3列的数据框。

Date          Close  pos_chg      cash
2000-01-04  81.9375      0.0  10000.00
2000-01-05  69.7500      0.0      0.00
2000-01-06  65.5625      0.0      0.00
2000-01-07  69.5625      0.0      0.00
2000-01-10  69.1875    100.0      0.00
2000-01-11  66.7500      0.0      0.00
2000-01-12  63.5625   -200.0      0.00
2000-01-13  65.9375    200.0      0.00
2000-01-14  64.2500   -200.0      0.00
2000-01-18  64.1250      0.0      0.00

我想要计算什么。

Date          Close  pos_chg      cash
2000-01-04  81.9375      0.0  10000.00
2000-01-05  69.7500      0.0  10000.00
2000-01-06  65.5625      0.0  10000.00
2000-01-07  69.5625      0.0  10000.00
2000-01-10  69.1875    100.0   3081.25
2000-01-11  66.7500      0.0   3081.25
2000-01-12  63.5625   -200.0  15793.75
2000-01-13  65.9375    200.0   2606.25
2000-01-14  64.2500   -200.0  15456.25
2000-01-18  64.1250      0.0  15456.25

我认为这段代码会起作用。

df['cash']=df['cash'].shift(1)-(df['pos_chg']*df['Close'])

任何帮助赞赏。

Answer 1

试试pandas.Series.cumsum()：

Code:

df['delta_cash'] = (df.Close * -df.pos_chg + df.cash).cumsum()

Test Code:

import pandas as pd
import numpy as np

df = pd.read_fwf(StringIO(u"""
    Date          Close  pos_chg      cash
    2000-01-04  81.9375      0.0  10000.00
    2000-01-05  69.7500      0.0      0.00
    2000-01-06  65.5625      0.0      0.00
    2000-01-07  69.5625      0.0      0.00
    2000-01-10  69.1875    100.0      0.00
    2000-01-11  66.7500      0.0      0.00
    2000-01-12  63.5625   -200.0      0.00
    2000-01-13  65.9375    200.0      0.00
    2000-01-14  64.2500   -200.0      0.00
    2000-01-18  64.1250      0.0      0.00"""),
                 skiprows=1)

df['delta_cash'] = (df.Close * -df.pos_chg + df.cash).cumsum()
print(df)

Results:

         Date    Close  pos_chg     cash  delta_cash
0  2000-01-04  81.9375      0.0  10000.0    10000.00
1  2000-01-05  69.7500      0.0      0.0    10000.00
2  2000-01-06  65.5625      0.0      0.0    10000.00
3  2000-01-07  69.5625      0.0      0.0    10000.00
4  2000-01-10  69.1875    100.0      0.0     3081.25
5  2000-01-11  66.7500      0.0      0.0     3081.25
6  2000-01-12  63.5625   -200.0      0.0    15793.75
7  2000-01-13  65.9375    200.0      0.0     2606.25
8  2000-01-14  64.2500   -200.0      0.0    15456.25
9  2000-01-18  64.1250      0.0      0.0    15456.25