Sql 在GROUPBY子句中使用min和max

我想要以下oracle sql中的输出

我的数据如下表所示：

id    start_date      end_date      assignment number

1     2.02.2014       15.02.2014    10
2     25.02.2014      30.02.2014    20
3     26.03.2014      04.05.2014    30
4     06.06.2014      31.12.4712    10

我需要使用分组输出

assignment_number   start_date          end_date

10                  02.02.2014          15.02.2014
10                  06.06.2014          31.12.4712
20                  25.02.2014          30.02.2014
30                  26.03.2014          04.05.2014

我试着使用最小值（开始日期）和最大值（结束日期）作为任务10的输出

assignment_number   start_date          end_date
10                  02.02.2014          31.12.4712

但我想：-

assignment_number   start_date          end_date
10                  02.02.2014          15.02.2014
10                  06.06.2014          31.12.4712

---

请帮忙

我想你必须分别计算最小值和最大值，然后合并它们。试着这样做：

SELECT 
 assignment_number
 , start_date
 , end_date
 FROM 
      (SELECT
                assignment_number
                , start_date
                , end_date
                FROM TABLE
                GROUP BY assignment_number
                HAVING MIN(start_date)

           UNION

           SELECT
                assignment_number
                , start_date
                , end_date
                FROM TABLE
                GROUP BY assignment_number
                HAVING MAX(end_date)
      )
 ORDER BY 
      1 ASC
      , 2 ASC
      , 3 ASC

)

---

就目前而言，您的需求没有真正意义。您需要为每个赋值指定min（）和max（）（在这种情况下，您的结果是正确的），或者需要表格中的每一行（在这种情况下，您根本不需要

分组依据

）。请提供有关预期输出的更多信息。

select id, to_char(start_date,'dd.mm.yyyy') start_date, to_char(end_date,'dd.mm.yyyy') end_date,ASSIGNMENT_NUMBER from sof1 s
where not exists 
                (select 1 from sof1 s2 
                 where s2.assignment_number=s.assignment_number 
                 and s2.start_date<s.start_date
                )
or not exists 
                (select 1 from sof1 s2 
                 where s2.assignment_number=s.assignment_number 
                 and s2.end_date>s.end_date
                )
order by ASSIGNMENT_NUMBER 

select id, to_char(start_date,'dd.mm.yyyy') start_date, to_char(end_date,'dd.mm.yyyy') end_date,ASSIGNMENT_NUMBER from 

(select s.*
 , min (start_date) over (partition by ASSIGNMENT_NUMBER) sd
 , max (end_date) over (partition by ASSIGNMENT_NUMBER) ed
 from sof1 s
)
where start_date=sd or end_date=ed
order by ASSIGNMENT_NUMBER, start_date