基于两个字段对评论进行分组

Question

有人要求我根据结束日期和 CD_TYPE_COMMENTS 字段对 TX_Comments 进行分组。输出应该是每个 CD_TYPE_COMMENTS 和 Enddate

的最新行

ID_EMPLOYEE ID_HEALTH_ASSESSMENT         CD_TYPE_COMMENT    ENDDATE TX_COMMENT
0M00910044  37754   Validation          18.06.2013          Record validated.
0M00910044  37754   Validation          20.05.2013          Record validated.
0M00910044  37754   Result              26.07.2013          Created.
0M00910044  37754   Appointment         18.06.2013          Pls cancel due to                        work commitments.  
0M00910044  37754   Appointment         25.06.2013          Confirmed (via Import)
0M00910044  37754   Ellipse Difference  23.07.2013  The following Employee Details have been updated:
0M00910044  37754   General             26.07.2013  Record complete
0M00910044  37754   Validation          20.05.2013  Record created.
0M00910044  37754   Appointment         09.07.2013  As advised by Mick Maskell. To be rescheduled to another date.
0M00910044  37754   Appointment         09.07.2013  Late cancellation charges will apply.
0M00910044  37754   Appointment         09.07.2013  Confirmed (via Import)
0M00910044  37754   Ellipse Difference  23.07.2013  The following Employee Details have been updated and should be checked for impact on this HA:
0M00910044  37754   Validation          20.05.2013  Manually created due to import exception error - Test Type could not be determined. Create HA manually. Has a valid RISI.
0M00910044  37754   Appointment         25.06.2013  Preferences provided.
0M00910044  37754   Ellipse Difference  04.07.2013  The following Employee Details have been updated and should be checked for impact on this HA:
0M00910044  37754   Validation          09.07.2013  Record validated.
0M00910044  37754   Appointment         18.06.2013  Cancelled (by Railcorp)
0M00910044  37754   Appointment         09.07.2013  Cancelled (by Railcorp)
0M00910044  37754   Appointment         24.05.2013  Preferences provided.
0M00910044  37754   Appointment         28.05.2013  Confirmed (via Import)
0M00910044  37754   Appointment         09.07.2013  Preferences provided.

Answer 1

如果您想要最后一行，那么您不想聚合数据。你只想select最后一个。

这是一种方法：

select t.*
from t
where t.enddate = (select max(t2.enddate)
                   from t t2
                   where t2.CD_TYPE_COMMENT = t.CD_TYPE_COMMENT
                  );

如果您确实还需要每个员工的此信息，请在关联子句中包含 and t2.Employee_ID = t.Employee_ID。

基于两个字段对评论进行分组

Group Comments based on two fields

sql

greatest-n-per-group