Python Pandas: 用算术前向填充

Question

我有以下数据框：

            AAPL shares  GOOG shares  MSFT shares
date                                             
2019-01-01          NaN         10.0          NaN
2019-01-05          NaN          NaN         15.0
2019-01-12          NaN          NaN          7.0
2019-01-13          3.0          NaN          NaN
2019-01-14          NaN         -5.0          NaN

应用前向填充后：

print(df.set_index('date').sort_index().fillna(method='ffill').fillna(value=0))

我得到：

            AAPL shares  GOOG shares  MSFT shares
date                                             
2019-01-01          0.0         10.0          0.0
2019-01-05          0.0         10.0         15.0
2019-01-12          0.0         10.0          7.0
2019-01-13          3.0         10.0          7.0
2019-01-14          3.0         -5.0          7.0

我的问题是有什么方法可以用简单的加法来填充前向吗？我要找的结果：

            AAPL shares  GOOG shares  MSFT shares
date                                             
2019-01-01          0.0         10.0          0.0
2019-01-05          0.0         10.0         15.0
2019-01-12          0.0         10.0         22.0
2019-01-13          3.0         10.0         22.0
2019-01-14          3.0          5.0         22.0

Answer 1

您可以查看 cumsum

df.fillna(0).cumsum()
            AAPLshares  GOOGshares  MSFTshares
date                                          
2019-01-01         0.0        10.0         0.0
2019-01-05         0.0        10.0        15.0
2019-01-12         0.0        10.0        22.0
2019-01-13         3.0        10.0        22.0
2019-01-14         3.0         5.0        22.0