如何在一分钟内计算第一秒和随后每一秒的增量

Question

我有一个csv文件，看起来像这样(大约400000行(它每秒钟从晚上10点到19点，它们被复制了很多

时间10:00:00610:00:0010:00:0010:00:00410:00:01610:00:01210:00:01910:00:04410:00:04510:00:041

Answer 1

IIUC，

import pandas as pd
df = pd.DataFrame([
['10:00:00',    6],
['10:00:00',    5],
['10:00:00',    2],
['10:00:00',    4],
['10:00:01',    6],
['10:00:01',    2],
['10:00:01',    9],
['10:00:04',    4],
['10:00:04',    5],
['10:00:04',    1],
['10:01:00',    1],
], columns=['time', 'value'])
data = df[pd.to_datetime(df.time,format= '%H:%M:%S' ).dt.second == 0].groupby('time').agg("first").reset_index()
data["hour_and_min"] = pd.to_datetime(data.time,format= '%H:%M:%S' ).dt.strftime("%H:%M")
res = df.groupby('time').agg("last").reset_index()
res["hour_and_min"] = pd.to_datetime(res.time,format= '%H:%M:%S' ).dt.strftime("%H:%M")
pd.DataFrame(pd.merge(res, data, on="hour_and_min").apply(lambda x: [x.value_x - x.value_y, f"{x.time_y}-{x.time_x}"], axis=1).tolist(), columns=['value', 'time'])

输出：

value   time
0   -2  10:00:00-10:00:00
1   3   10:00:00-10:00:01
2   -5  10:00:00-10:00:04