如何使用Oracle SQL为大多数当前部门的员工查找特定日期？

Question

我试图确定在其最近部门的记录（见下面的数据）雇员的最小有效日期。 我期待有2014年8月25日的结果，而是想从你一些专家协助那里就如何把它关闭。 在相同的数据，该员工在部门70260一会儿，然后转移时间以70210短的一段时间，然后回来70260.的2014年8月25日生效日期为当日的员工回来和我倒要看看是否有编程SQL显示某人的最小日期在其最近部门的任何创造性的方式。 提前致谢！！！

╔═══════════╦════════════╦════════╗
║  EMPLID   ║   EFFDT    ║ DEPTID ║
╠═══════════╬════════════╬════════╣
║ 000123338 ║ 10/25/2015 ║  70260 ║
║ 000123338 ║ 4/2/2015   ║  70260 ║
║ 000123338 ║ 2/24/2015  ║  70260 ║
║ 000123338 ║ 11/1/2014  ║  70260 ║
║ 000123338 ║ 8/25/2014  ║  70260 ║ <--- I need the SQL to show 8/25/2014
║ 000123338 ║ 4/27/2014  ║  70210 ║
║ 000123338 ║ 3/16/2014  ║  70210 ║
║ 000123338 ║ 3/6/2014   ║  70260 ║
║ 000123338 ║ 11/1/2013  ║  70260 ║
║ 000123338 ║ 1/24/2013  ║  70260 ║
║ 000123338 ║ 1/1/2013   ║  70260 ║
║ 000123338 ║ 11/1/2012  ║  70260 ║
║ 000123338 ║ 11/1/2011  ║  70260 ║
║ 000123338 ║ 8/1/2010   ║  70260 ║
║ 000123338 ║ 8/5/2009   ║  70260 ║
║ 000123338 ║ 7/1/2009   ║  70260 ║
║ 000123338 ║ 7/7/2008   ║  70260 ║
║ 000123338 ║ 5/5/2008   ║  70260 ║
║ 000123338 ║ 1/1/2008   ║  70260 ║
║ 000123338 ║ 10/29/2007 ║  70260 ║
╚═══════════╩════════════╩════════╝

Answer 1

你可以做这样的事情

select * from 
(select empid,effdt,deptid,
 row_number() over ( partition by deptid order by effdt desc) as mostRecent
from yourtable) a
where a.mostRecent=1

Answer 2

您可以结合使用窗口聚合功能和最终的全局聚合：

select   empid,
         deptid,
         min(effdt) min_effdt
from     (select empid,
                 deptid,
                 row_number() over 
                     (partition by empid order by effdt desc) as rn1,
                 row_number() over 
                     (partition by empid, deptid order by effdt desc) as rn2
          from   tbl) t
where    rn1 = rn2
group by empid,
         deptid

内部查询返回这些行号，而外部查询从其余记录（具有RN1 = RN2）中获取最小日期。

╔═══════════╦════════════╦════════╦═════╦═════╦═══════════╗
║  EMPLID   ║   EFFDT    ║ DEPTID ║ RN1 ║ RN2 ║ MIN_EFFDT ║ 
╠═══════════╬════════════╬════════╬═════╬═════╬═══════════╣
║ 000123338 ║ 10/25/2015 ║  70260 ║  1  ║  1  ║           ║  
║ 000123338 ║  4/02/2015 ║  70260 ║  2  ║  2  ║           ║  
║ 000123338 ║  2/24/2015 ║  70260 ║  3  ║  3  ║           ║  
║ 000123338 ║ 11/01/2014 ║  70260 ║  4  ║  4  ║           ║  
║ 000123338 ║  8/25/2014 ║  70260 ║  5  ║  5  ║ 8/25/2014 ║  
║ 000123338 ║  4/27/2014 ║  70210 ║  6  ║  1  ╠═══════════╝  
║ 000123338 ║  3/16/2014 ║  70210 ║  7  ║  2  ║  
║ 000123338 ║  3/06/2014 ║  70260 ║  8  ║  6  ║  
║ 000123338 ║ 11/01/2013 ║  70260 ║  9  ║  7  ║  
║ 000123338 ║  1/24/2013 ║  70260 ║ 10  ║  8  ║  
║ 000123338 ║  1/01/2013 ║  70260 ║ 11  ║  9  ║  
║ 000123338 ║ 11/01/2012 ║  70260 ║ 12  ║ 10  ║  
║ 000123338 ║ 11/01/2011 ║  70260 ║ 13  ║ 11  ║  
║ 000123338 ║  8/01/2010 ║  70260 ║ 14  ║ 12  ║  
║ 000123338 ║  8/05/2009 ║  70260 ║ 15  ║ 13  ║  
║ 000123338 ║  7/01/2009 ║  70260 ║ 16  ║ 14  ║  
║ 000123338 ║  7/07/2008 ║  70260 ║ 17  ║ 15  ║  
║ 000123338 ║  5/05/2008 ║  70260 ║ 18  ║ 16  ║  
║ 000123338 ║  1/01/2008 ║  70260 ║ 19  ║ 17  ║  
║ 000123338 ║ 10/29/2007 ║  70260 ║ 20  ║ 18  ║  
╚═══════════╩════════════╩════════╩═════╩═════╝

Answer 3

您需要找到第一组行：

SELECT EMPLID, MIN(first_dates)
FROM
 (
    SELECT EMPLID, EFFDT, DEPTID,
       CASE -- check if all previous rows return the same value
          WHEN MIN(DEPTID) OVER (PARTITION BY EMPLID ORDER BY EFFDT DESC ROWS UNBOUNDED PRECEDING)
             = MAX(DEPTID) OVER (PARTITION BY EMPLID ORDER BY EFFDT DESC ROWS UNBOUNDED PRECEDING)
          THEN EFFDT
       END AS first_dates -- all other rows return NULL
    FROM tab
 ) dt
GROUP BY 1;

如果要整行，则需要添加另一个ROW_NUMBER：

SELECT ...
FROM
 (
   SELECT ...,
      ROW_NUMBER() OVER (PARTITION BY EMPLID ORDER BY first_dates) AS rn
   FROM
    (
       SELECT EMPLID, EFFDT, DEPTID,
          CASE
             WHEN MIN(DEPTID) OVER (PARTITION BY EMPLID ORDER BY EFFDT DESC ROWS UNBOUNDED PRECEDING)
                = MAX(DEPTID) OVER (PARTITION BY EMPLID ORDER BY EFFDT DESC ROWS UNBOUNDED PRECEDING)
             THEN EFFDT
          END AS first_dates
       FROM tab
    ) dt
 ) dt
WHERE rn = 1