Question

我有两个表通过一个简单的键链接，一个包含项目名称及其开始日期，另一个包含规划，如下所示：

projects
ID      NAME  START_DATE
1       foo   01/01/2017
2       barr  01/02/2017

planning
PRJ    M0    M1    M2
1      70    75    80
2      50    60    70

有人可以帮我加入这两张桌子以便制作：

PRJ    DATE        PLAN
1      01/01/2017  70
1      01/02/2017  75
1      01/03/2017  80
2      01/02/2017  50
2      01/03/2017  60
2      01/04/2017  70

unpivot似乎在这里很有用，但我没有多少练习。

感谢

Answer 1

考虑在CTE中运行UNPIVOT。然后，在主查询中计算可用于将日期添加到 START_DATE 的行号。

另外，请考虑不要使用M*列以宽格式存储数据，而应使用长格式的原始值和指示符列。否则，下面的查询将很复杂。

WITH master AS
  (SELECT *
   FROM planning
   UNPIVOT ("PLAN" FOR M_COLS IN (M0, M1, M2))  -- BUILD OUT LIST IN APP CODE
  )

SELECT m.PRJ, p.START_DATE + m.row_num AS m."DATE", m."PLAN"
FROM projects p
INNER JOIN 
   (SELECT PRJ, "PLAN", ROW_NUMBER() 
                        OVER (PARTITION BY PRJ ORDER BY PRJ) - 1 AS row_num
    FROM master) AS m
ON p.ID = m.PRJ;

Answer 2

当您取消忽略时，跟踪不透明列的列可以是任何数据类型。在这种情况下，您可以将其设为NUMBER列，并且可以将值0, 1, 2, ...分配给列M0, M1, M2, ... - 这样您就可以直接使用它来调整开始日期。请参阅IN子句的UNPIVOT列表。

with
  projects ( id, name, start_date ) as (
    select 1, 'foo', to_date('01/01/2017', 'mm/dd/yyyy') from dual union all
    select 2, 'bar', to_date('01/02/2017', 'mm/dd/yyyy') from dual
  ),
  planning ( prj, m0, m1, m2 ) as (
    select 1, 70, 75, 80 from dual union all
    select 2, 50, 60, 70 from dual
  )
-- End of simulated inputs (for testing only, not part of the solution).
-- SQL query begins BELOW THIS LINE.
select x.prj, p.start_date + x.col_num as start_date, x.plan
from   projects p
       join
       ( select prj, col_num, plan
         from   planning
         unpivot ( plan for col_num in (m0 as 0, m1 as 1, m2 as 2) )
       ) x
             on p.id = x.prj
order by prj, start_date      -- If needed.
;

       PRJ START_DATE       PLAN
---------- ---------- ----------
         1 01/01/2017         70
         1 01/02/2017         75
         1 01/03/2017         80
         2 01/02/2017         50
         2 01/03/2017         60
         2 01/04/2017         70

Oracle在添加日期时忽略了

2 个答案: