Pandas Resample With Overlap

I would like to resample my time indexed DataFrame with some overlapping. For example: >>> df data date 2018-03-09 12:00:00 1

Solution 1:

Assuming your data are regularly spaced you can concat the shifted Series and then take a sum.

N = 10  # Every 10 seconds from first row
ov = 2  # 2s overlap on either sidepd.concat([df.shift(i).iloc[::N] for i in range(-ov, N+ov)], axis=1).sum(1)

---

date2018-03-09 12:00:00     3.02018-03-09 12:00:10    15.02018-03-09 12:00:20    28.02018-03-09 12:00:30    34.0dtype:float64

---

For multiple columns, turn this into a groupby along the columns axis:

df['data2'] = df['data']+1  # Another column

(pd.concat([df.shift(i).iloc[::N] for i in range(-ov, N+ov)], axis=1)
   .groupby(level=0, axis=1).sum())

                     data  data2
date                            
2018-03-09 12:00:00   3.0    6.0
2018-03-09 12:00:10  15.0   28.0
2018-03-09 12:00:20  28.0   42.0
2018-03-09 12:00:30  34.0   46.0