Python 使用特定的列值在 pandas 中保留具有真实重叠时间段的行

您可以对时间进行排序，然后使用groupby.shift比较boolean indexing的连续行:

# ensure datetime
df[['Start DateTime', 'End DateTime']] = df[['Start DateTime', 'End DateTime']].apply(pd.to_datetime)

# sort values
tmp = df.sort_values(by=['values_1', 'Start DateTime', 'End DateTime'])

# are consecutive intervals overlapping?
m = tmp['Start DateTime'].lt(tmp.groupby('values_1')['End DateTime'].shift())

# flag rows and previous ones
out = df.loc[m|m.groupby(tmp['values_1']).shift(-1)]

输出:

  values_1  uniqId DeptId      Start DateTime        End DateTime
0      VGM    1278   BKNG 2023-09-05 18:35:28 2023-09-05 20:05:28
1      VGM    1259   BKNG 2023-09-05 18:55:18 2023-09-05 19:25:38
4      WEH    9124   ASKG 2023-09-05 17:25:28 2023-09-05 19:43:13
5      WEH    9785   ASKG 2023-09-05 16:15:21 2023-09-05 18:24:02

中间体:

  values_1  uniqId DeptId      Start DateTime        End DateTime           end_shift      m m_shift  m|m_shift
6      POH    3479   ASKG 2023-09-05 15:35:29 2023-09-05 17:25:22                 NaT  False   False      False
7      POH    3449   ASKG 2023-09-05 17:35:28 2023-09-05 18:35:19 2023-09-05 17:25:22  False     NaN      False
2      VGM    2567   BKNG 2023-09-05 14:29:38 2023-09-05 17:35:28                 NaT  False   False      False
0      VGM    1278   BKNG 2023-09-05 18:35:28 2023-09-05 20:05:28 2023-09-05 17:35:28  False    True       True
1      VGM    1259   BKNG 2023-09-05 18:55:18 2023-09-05 19:25:38 2023-09-05 20:05:28   True     NaN       True
3      WEH    9854   ASKG 2023-09-05 13:45:58 2023-09-05 15:00:00                 NaT  False   False      False
5      WEH    9785   ASKG 2023-09-05 16:15:21 2023-09-05 18:24:02 2023-09-05 15:00:00  False    True       True
4      WEH    9124   ASKG 2023-09-05 17:25:28 2023-09-05 19:43:13 2023-09-05 18:24:02   True     NaN       True