Question

我有一个像这样的ORACLE表

---------------------------------------
DATE        |  USERID | DOMAIN | VALUE
---------------------------------------
03/16/2016     1001     ASIA      10
03/16/2016     1001     EUROPE    20
03/16/2016     1002     ASIA      20
03/17/2016     1001     ASIA      20
03/17/2016     1002     EUROPE    10
----------------------------------------

我想将此表格翻译成类似

的视图

 -------------------------------------
 DATE       |  USERID | ASIA | EUROPE  
 -------------------------------------
 03/16/2016     1001     10     20
 03/16/2016     1002     20      
 03/17/2016     1001     20
 03/17/2016     1002            10 
 -------------------------------------

如果我尝试使用PIVOT，我可以在用户级别上进行，但不知道如何获取日期，用户级别。任何指针都会很棒。

Answer 1

PIVOT似乎达到了你需要的结果：

SQL> with test (DATE_, USERID, DOMAIN, VALUE)
  2  as (
  3   select '03/16/2016',     1001     ,'ASIA'      ,10 from dual union all
  4   select '03/16/2016',     1001     ,'EUROPE'    ,20 from dual union all
  5   select '03/16/2016',     1002     ,'ASIA'      ,20 from dual union all
  6   select '03/17/2016',     1001     ,'ASIA'      ,20 from dual union all
  7   select '03/17/2016',     1002     ,'EUROPE'    ,10 from dual
  8  )
  9  SELECT *
 10  FROM   (select *
 11          from test)
 12  PIVOT ( sum(value) FOR (domain) IN ('ASIA', 'EUROPE'))
 13  ORDER BY 1, 2;

DATE_          USERID     'ASIA'   'EUROPE'
---------- ---------- ---------- ----------
03/16/2016       1001         10         20
03/16/2016       1002         20
03/17/2016       1001         20
03/17/2016       1002                    10

Answer 2

执行->addselect('sizes')，使用GROUP BY表达式选择亚洲或欧洲：

case

回答不支持现代Oracle版本的产品＆＃39; select DATE, USERID, sum(case when DOMAIN = 'ASIA' then VALUE end) as asia sum(case when DOMAIN = 'EUROPE' then VALUE end) as europe from tablename group by DATE, USERID。

Answer 3

您可以使用PIVOT功能

pivot 
(
   count(VALUE)
   for DOMAIN in ("ASIA", "EUROPE")
)

，将查询包装在子选择中，然后按用户ID对结果进行分组。

根据行值多级层次结构将行值转换为列

3 个答案: