我尝试使用以下csv数据获取经过的时间，但是当日期交叉时，Elapsed_time中的秒数显示为减号。

import numpy as np
import datetime as dt
import matplotlib.pyplot as plt
import pylab as pl
import pandas as pd 

dft = pd.read_csv("/content/drive/MyDrive/toa_data/to_csv_out_columns.csv", encoding="UTF-8")
dft["Time"] = pd.to_datetime(dft["Time"])
dft["second"] = pd.to_timedelta(dft["Time"].dt.time.astype(str)).dt.total_seconds()#.div(60)

dft = dft.set_index("second")
#dft.index = pd.to_datetime(dft.index)
dft['diff'] = dft.index.to_series().diff()
dft["Elapsed_time"]=dft["diff"].cumsum()

dft.head(5)

second  Time    Elapsed_time    value   diff
72446   2023-04-22 20:07:26     3.16    
72506   2023-04-22 20:08:26 60  3.16    60
72566   2023-04-22 20:09:26 120 3.17    60
72626   2023-04-22 20:10:26 180 3.18    60
--------------omission-----------------------------------
86306   2023-04-22 23:58:26 13860   3.41    60
86366   2023-04-22 23:59:26 13920   3.41    60
26  2023-04-23 0:00:26  -72420  3.41    -86340
86  2023-04-23 0:01:26  -72360  3.41    60
146 2023-04-23 0:02:26  -72300  3.41    60
--------------------------------------------------

我想输出以下csv。

second  Time    Elapsed_time    value   diff
72446   2023-04-22 20:07:26     3.16    
72506   2023-04-22 20:08:26 60  3.16    60
72566   2023-04-22 20:09:26 120 3.17    60
72626   2023-04-22 20:10:26 180 3.18    60
--------------omission----------------------------
86306   2023-04-22 23:58:26 13860   3.41    60
86366   2023-04-22 23:59:26 13920   3.41    60
86426   2023-04-23 0:00:26  13960   3.41    60
86486   2023-04-23 0:01:26  14020   3.41    60
86546   2023-04-23 0:02:26  14080   3.41    60
--------------------------------------------------