如果为 NaN，则滚动数据框 pandas 并更新值

Question

I want to use rolling function on my dataframe. Here my series :

index	date	id	value_dex
90256	2021-05-10 01:45:20	101904285	7.6
90257	2021-05-10 01:45:20	101904285	7.6
90258	2021-05-10 02:00:00	101904285	NaN
90260	2021-05-10 02:00:44	101904285	6.9
90261	2021-05-10 02:00:44	101904285	NaN

What I want :

index	date	id	value_dex
90256	2021-05-10 01:45:20	101904285	7.6
90257	2021-05-10 01:45:20	101904285	7.6
90258	2021-05-10 02:00:00	101904285	7.6
90260	2021-05-10 02:00:44	101904285	6.9
90261	2021-05-10 02:00:44	101904285	6.9

I would like to drag the values in the column over 30 minutes if NaN. However, I also want the dragging to stop when a new value is encountered on the 30 minute range (in this case, we drag the new value over 30 minutes).

I tried that :

ffill = lambda x: pd.DataFrame(x.rolling('30T', on='Date')[y].max(),
                                       index=x.index)
        
        data[y] = data.groupby('ID', as_index=False).apply(ffill)

But it doesn't work when there is a new value during the rolling. Do you have any solutions in order to resolve my problem ? Maybe an existing function ? Thanks a lot !

Answer 1

我相信您可以尝试将其用作 NaN 值的填充物：

data.fillna(method='ffill')