每小时选择最新员工条目的 Oracle SQL 查询答案

【问题标题】：每小时选择最新员工条目的 Oracle SQL 查询
【发布时间】：2021-11-26 22:22:00
【问题描述】：

我将如何修改以下 SQL 查询代码以每小时返回每个员工的最新条目，而不是当前返回每个员工的每个事件，无论它每小时在查询中出现多少次。

我已附上示例数据，以说明当前在数据库中的顺序和所需的表输出。注意：您可以看到它是按降序排列的，并且省略了 2 个条目（0001 和 0009），因为它们在一个小时内来自同一员工。

SQL查询代码：

select
     TRANSACTION_ID
     EMPLOYEE_ID
     FIRST_NAME
     LAST_NAME
     TIME_STAMP
from (select d.*,
             row_number() over (partition by EMPLOYEE_ID, trunc(TIME_STAMP, 'HH') order by TIME_STAMP desc) as SEQNUM
      from MAIN.DATABASE d
     ) d
where SEQNUM = 1;
order by TRANSACTION_ID desc;

这是数据库中数据的当前顺序：

TRANSACTION_ID	EMPLOYEE_ID	FIRST_NAME	LAST_NAME	TIME_STAMP
0001	AAAA	Adam	Akbar	10/05/2021 04:42:42.000 PM
0004	BBBB	Barry	Brink	10/06/2021 07:25:25.000 AM
0003	CCCC	Charlie	Che	10/06/2021 07:15:15.000 AM
0005	DDDD	David	Doe	10/06/2021 07:27:27.000 AM
0006	EEEE	Eric	Erickson	10/06/2021 07:29:29.000 AM
0007	FFFF	Fred	Foe	10/06/2021 07:31:31.000 AM
0008	GGGG	George	Guy	10/06/2021 07:33:33.000 AM
0010	HHHH	Henry	Hugh	10/06/2021 07:55:55.000 AM
0009	HHHH	Henry	Hugh	10/06/2021 07:54:54.000 AM
0002	AAAA	Adam	Akbar	10/05/2021 04:43:43.000 PM

这是应该返回的：

TRANSACTION_ID	EMPLOYEE_ID	FIRST_NAME	LAST_NAME	TIME_STAMP
0010	HHHH	Henry	Hugh	10/06/2021 07:55:55.000 AM
0008	GGGG	George	Guy	10/06/2021 07:33:33.000 AM
0007	FFFF	Fred	Foe	10/06/2021 07:31:31.000 AM
0006	EEEE	Eric	Erickson	10/06/2021 07:29:29.000 AM
0005	DDDD	David	Doe	10/06/2021 07:27:27.000 AM
0004	BBBB	Barry	Brink	10/06/2021 07:25:25.000 AM
0003	CCCC	Charlie	Che	10/06/2021 07:15:15.000 AM
0002	AAAA	Adam	Akbar	10/05/2021 04:43:43.000 PM

但是，这是代码当前返回的内容：

TRANSACTION_ID	EMPLOYEE_ID	FIRST_NAME	LAST_NAME	TIME_STAMP
0010	HHHH	Henry	Hugh	10/06/2021 07:55:55.000 AM
0006	EEEE	Eric	Erickson	10/06/2021 07:29:29.000 AM
0003	CCCC	Charlie	Che	10/06/2021 07:15:15.000 AM
0002	AAAA	Adam	Akbar	10/05/2021 04:43:43.000 PM

知道我缺少什么以及如何解决它吗？

【问题讨论】：

标签： sql oracle greatest-n-per-group

【解决方案1】：

查询（为了返回你想要的）应该这样工作。但是，我假设您的字段 time_stamp 确实是您提供的格式的时间戳。

SQL> set lines 220
SQL> with x ( transaction_id , employee_id, first_name, last_name, time_stamp )
  2  as
  3  (
  4  select '0001' , 'AAAA', 'Adam'       , 'Akbar'     , to_timestamp('10/05/2021 04:42:42.000 PM','MM/DD/YYYY HH:MI:SS.FF3 PM') from dual union all
  5  select '0004' , 'BBBB', 'Barry'      , 'Brink'     , to_timestamp('10/06/2021 07:25:25.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
  6  select '0003' , 'CCCC', 'Charlie' , 'Che'      , to_timestamp('10/06/2021 07:15:15.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
  7  select '0005' , 'DDDD', 'David'      , 'Doe'           , to_timestamp('10/06/2021 07:27:27.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
  8  select '0006' , 'EEEE', 'Eric'       , 'Erickson'  , to_timestamp('10/06/2021 07:29:29.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
  9  select '0007' , 'FFFF', 'Fred'       , 'Foe'           , to_timestamp('10/06/2021 07:31:31.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
 10  select '0008' , 'GGGG', 'George'  , 'Guy'      , to_timestamp('10/06/2021 07:33:33.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
 11  select '0010' , 'HHHH', 'Henry'      , 'Hugh'          , to_timestamp('10/06/2021 07:55:55.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
 12  select '0009' , 'HHHH', 'Henry'      , 'Hugh'          , to_timestamp('10/06/2021 07:54:54.000 AM','MM/DD/YYYY HH:MI:SS.FF3 AM') from dual union all
 13  select '0002' , 'AAAA', 'Adam'       , 'Akbar'         , to_timestamp('10/05/2021 04:43:43.000 PM','MM/DD/YYYY HH:MI:SS.FF3 PM') from dual
 14  )
 15  select
 16       TRANSACTION_ID ,
 17       EMPLOYEE_ID    ,
 18       FIRST_NAME     ,
 19       LAST_NAME      ,
 20       TIME_STAMP
 21  from (select x.*,
 22               row_number() over (partition by EMPLOYEE_ID, trunc(TIME_STAMP, 'HH') order by TIME_STAMP desc) as SEQNUM
 23        from x
 24       )
 25  where SEQNUM = 1
 26* order by TRANSACTION_ID desc

TRAN EMPL FIRST_N LAST_NAM TIME_STAMP
---- ---- ------- -------- ---------------------------------------------------------------------------
0010 HHHH Henry   Hugh     06-OCT-21 07.55.55.000000000 AM
0008 GGGG George  Guy      06-OCT-21 07.33.33.000000000 AM
0007 FFFF Fred    Foe      06-OCT-21 07.31.31.000000000 AM
0006 EEEE Eric    Erickson 06-OCT-21 07.29.29.000000000 AM
0005 DDDD David   Doe      06-OCT-21 07.27.27.000000000 AM
0004 BBBB Barry   Brink    06-OCT-21 07.25.25.000000000 AM
0003 CCCC Charlie Che      06-OCT-21 07.15.15.000000000 AM
0002 AAAA Adam    Akbar    05-OCT-21 04.43.43.000000000 PM

8 rows selected.

SQL>

【讨论】：

【解决方案2】：

您的查询是：

SELECT 子句中的术语之间缺少逗号；和
在WHERE 过滤器之后和ORDER BY 子句之前有一个;。

如果你解决了这些问题，那么你会得到代码：

select TRANSACTION_ID,
       EMPLOYEE_ID,
       FIRST_NAME,
       LAST_NAME,
       TIME_STAMP
from   (
  select d.*,
         row_number() over (
           partition by EMPLOYEE_ID, trunc(TIME_STAMP, 'HH')
           order by TIME_STAMP desc
         ) as SEQNUM
  from   MAIN.DATABASE d
) d
where  SEQNUM = 1
order by TRANSACTION_ID desc;

其中，对于样本数据：

CREATE TABLE main.database (TRANSACTION_ID, EMPLOYEE_ID, FIRST_NAME, LAST_NAME, TIME_STAMP ) AS
SELECT '0001', 'AAAA', 'Adam',    'Akbar',    TIMESTAMP '2021-05-10 16:42:42.000' FROM DUAL UNION ALL
SELECT '0004', 'BBBB', 'Barry',   'Brink',    TIMESTAMP '2021-06-10 07:25:25.000' FROM DUAL UNION ALL
SELECT '0003', 'CCCC', 'Charlie', 'Che',      TIMESTAMP '2021-06-10 07:15:15.000' FROM DUAL UNION ALL
SELECT '0005', 'DDDD', 'David',   'Doe',      TIMESTAMP '2021-06-10 07:27:27.000' FROM DUAL UNION ALL
SELECT '0006', 'EEEE', 'Eric',    'Erickson', TIMESTAMP '2021-06-10 07:29:29.000' FROM DUAL UNION ALL
SELECT '0007', 'FFFF', 'Fred',    'Foe',      TIMESTAMP '2021-06-10 07:31:31.000' FROM DUAL UNION ALL
SELECT '0008', 'GGGG', 'George',  'Guy',      TIMESTAMP '2021-06-10 07:33:33.000' FROM DUAL UNION ALL
SELECT '0010', 'HHHH', 'Henry',   'Hugh',     TIMESTAMP '2021-06-10 07:55:55.000' FROM DUAL UNION ALL
SELECT '0009', 'HHHH', 'Henry',   'Hugh',     TIMESTAMP '2021-06-10 07:54:54.000' FROM DUAL UNION ALL
SELECT '0002', 'AAAA', 'Adam',    'Akbar',    TIMESTAMP '2021-05-10 16:43:43.000' FROM DUAL;

按预期工作：

TRANSACTION_ID EMPLOYEE_ID FIRST_NAME LAST_NAME TIME_STAMP

0010 HHHH Henry Hugh 10-JUN-21 07.55.55.000000000

0008 GGGG George Guy 10-JUN-21 07.33.33.000000000

0007 FFFF Fred Foe 10-JUN-21 07.31.31.000000000

0006 EEEE Eric Erickson 10-JUN-21 07.29.29.000000000

0005 DDDD David Doe 10-JUN-21 07.27.27.000000000

0004 BBBB Barry Brink 10-JUN-21 07.25.25.000000000

0003 CCCC Charlie Che 10-JUN-21 07.15.15.000000000

0002 AAAA Adam Akbar 10-MAY-21 16.43.43.000000000

TRANSACTION_ID	EMPLOYEE_ID	FIRST_NAME	LAST_NAME	TIME_STAMP
0010	HHHH	Henry	Hugh	10-JUN-21 07.55.55.000000000
0008	GGGG	George	Guy	10-JUN-21 07.33.33.000000000
0007	FFFF	Fred	Foe	10-JUN-21 07.31.31.000000000
0006	EEEE	Eric	Erickson	10-JUN-21 07.29.29.000000000
0005	DDDD	David	Doe	10-JUN-21 07.27.27.000000000
0004	BBBB	Barry	Brink	10-JUN-21 07.25.25.000000000
0003	CCCC	Charlie	Che	10-JUN-21 07.15.15.000000000
0002	AAAA	Adam	Akbar	10-MAY-21 16.43.43.000000000

db小提琴here

【讨论】：

您好 MT0，感谢您的回复。我做了这些更正，但是，查询返回 0 个条目。我意识到这是因为在这种情况下，正确条目的 SEQNUM = 2（对于我们想要丢弃的条目，SEQNUM = 3），但在其他数据集中，好的 SEQNUM 条目 = 1（而坏的 SEQNUM 条目 = 2）。您是否知道在此代码中合并 [MAX(SEQNUM) - 1] 变量以在任何情况下标准化此代码的方法？