从 df.rolling 返回错误的 window

Question

I'm converting a series of data from 1 minute intervals to 5 minute intervals. To do this I am using the rolling and sum funcitons from pandas then attempting to slice in steps of 5. This makes sense to me, sum everything up, then take the rows that have the information I want, which is every 5th row.

However my code is not slicing how I intended and is instead ignoring row 0 for the operations. In the attached picture the Left column is the sliced code and the right column is the unsliced code. As the Picture of Results shows, the first slice happens at row 5 instead of 4 (the 5th data entry). I am geting the right length back in my slice, but clearly not getting all of the data I need. My code is below with tmpL on the Left and tmpR on the right. I am expecting to get rows 4 and 9 back instead of rows 5 and 10. What is the proper notation I should be using?

tmpL = df.rolling(window=5).sum()[::5]
tmpR = df.rolling(window=5).sum()

Answer 1

If this was me I would take the following approach:

tmp_df = df.rolling(window=5).sum()
df_5 = tmp_df[(tmp_df.index % 5) == 4]

df_5 will be your desired output