Pandas Rolling Python to create new Columns

问题

I have a dataframe from excel file with data something like this:

     A                     B                        Sum A       Sum B
0                                           
1   235353.21333333332                      
2    89160.59999999999                      
3   188382.98666666663                      
4                        104677.1466666667  
5                       207723.25333333333  
6                       170128.02666666667  
7                                 165287.5  
8   44863.200000000004                      
9                                177096.72  
10   97687.71666666666                          655447.7167    824912.6467
11  113207.76333333334                          533302.2667    824912.6467
12                                195151.2      444141.6667    1020063.847
13                       151408.4433333333        255758.68     1171472.29
14                       50865.66999999999        255758.68    1117660.813 
15                       84536.19000000002        255758.68      994473.75
16  217555.28666666665                          473313.9667    824345.7233
17   90395.21666666666                          563709.1833    659058.2233
18           126856.21                          645702.1933    659058.2233
19  125190.61999999998                          770892.8133    481961.5033

In excel file, to get Sum A at row 10 it uses =SUM(A2:A12) and to get Sum B at row 10 it uses =SUM(B2:B11). Now, I am trying to recreate it by using .rolling but it cause DataError: No numeric types of aggregate. The following is my code to recreate Sum A and Sum B:

df['Sum A'] = df['A'].rolling(10).sum()
df['Sum C'] = df['B'].rolling(10).sum()

Any suggestion how should I do it in order to recreate the value in Sum A and Sum B? Thank you in advance

回答1:

Can you try:

df['Sum A'] = pd.to_numeric(df['A'], errors='coerce').fillna(0).rolling(10).sum()

来源：https://stackoverflow.com/questions/64019861/pandas-rolling-python-to-create-new-columns