主页 / coding / 问题 / 79182408
Accepted
Hakim
Asked: 2024-11-13 03:00:49 +0800 CST

将 Pandas 中的时间序列转换为间隔

  • 772

我有如下的时间序列：

    mode     timestamp               
0    A    2021-06-29 00:00:00-04:00
1    A    2021-06-29 01:00:00-04:00
2    A    2021-06-29 02:00:00-04:00
3    A    2021-06-29 03:00:00-04:00
4    B    2021-06-29 04:00:00-04:00
5    B    2021-06-29 05:00:00-04:00
6    B    2021-06-29 06:00:00-04:00
7    B    2021-06-29 07:00:00-04:00
8    B    2021-06-29 08:00:00-04:00
9    B    2021-06-29 09:00:00-04:00
10   B    2021-06-29 10:00:00-04:00
11   A    2021-06-29 11:00:00-04:00
12   A    2021-06-29 12:00:00-04:00
13   A    2021-06-29 13:00:00-04:00
14   A    2021-06-29 14:00:00-04:00
15   A    2021-06-29 15:00:00-04:00
16   A    2021-06-29 16:00:00-04:00
17   A    2021-06-29 17:00:00-04:00
18   A    2021-06-29 18:00:00-04:00
19   A    2021-06-29 19:00:00-04:00
20   C    2021-06-29 20:00:00-04:00
21   C    2021-06-29 21:00:00-04:00
22   C    2021-06-29 22:00:00-04:00
23   C    2021-06-29 23:00:00-04:00
24   C    2021-06-29 00:00:00-04:00
25   C    2021-06-29 01:00:00-04:00
26   C    2021-06-29 02:00:00-04:00
27   C    2021-06-29 03:00:00-04:00

我正在尝试生成一个带有模式间隔的字典列表，其中包含开始和结束时间。像这样

[
{"mode": "A", start_time: "2021-06-29 00:00:00-04:00", end_time:"2021-06-29 03:00:00-04:00" },
{"mode": "B", start_time: "2021-06-29 04:00:00-04:00", end_time:"2021-06-29 10:00:00-04:00" },
{"mode": "A", start_time: "2021-06-29 11:00:00-04:00", end_time:"2021-06-29 19:00:00-04:00" },
{"mode": "C", start_time: "2021-06-29 20:00:00-04:00", end_time:"2021-06-29 03:00:00-04:00" },
]

时间序列很长，循环它们似乎很慢。有什么建议可以实现这一点吗？

pandas

1 个回答

  1. Best Answer

    您可以使用自定义的groupby.aggand to_dict：

    # group successive values
group = df['mode'].ne(df['mode'].shift()).cumsum()

# aggregate and convert to dictionary
out = (df.groupby(group)
         .agg(**{'mode': ('mode', 'first'),
                 'start_time': ('timestamp', 'first'),
                 'end_time': ('timestamp', 'last')})
         .to_dict('records')
      )

    注意：根据所需的逻辑，使用min/max代替first/last作为时间戳。

    输出：

    [{'mode': 'A', 'start_time': '2021-06-29 00:00:00-04:00', 'end_time': '2021-06-29 03:00:00-04:00'},
 {'mode': 'B', 'start_time': '2021-06-29 04:00:00-04:00', 'end_time': '2021-06-29 10:00:00-04:00'},
 {'mode': 'A', 'start_time': '2021-06-29 11:00:00-04:00', 'end_time': '2021-06-29 19:00:00-04:00'},
 {'mode': 'C', 'start_time': '2021-06-29 20:00:00-04:00', 'end_time': '2021-06-29 03:00:00-04:00'}]
    • 2

相关问题