如何在没有过程的情况下使用日期范围连接表

Question

我有两个表，我们称它们为表A &B: -

表A如下所示:

   A.Num1       |   A.Num2      |   A.Date      
   12345        |   38170       |   28/05/2013      
   12345        |   38170       |   29/05/2013      
   12345        |   38170       |   31/05/2013      
   12345        |   38170       |   01/06/2013      
   12345        |   38170       |   03/06/2013      
   12345        |   38170       |   04/06/2013  
   12345        |   38170       |   04/06/2013          
   12345        |   38170       |   07/06/2013  

表2 (B)如下所示:

    B.Num1      |   B.Num2      |   B.Status    |   B.Date
    12345       |   38170       |   New         |   28/05/2013  
    12345       |   38170       |   Closed      |   31/05/2013  
    12345       |   38170       |   Reopened    |   04/06/2013

我需要一个这样的表作为输出-它基本上应该包含表A中的所有行+表B中的状态

    Num1        |   Num2        |   Status      |   Date
    12345       |   38170       |   New         |   28/05/2013      
    12345       |   38170       |   New         |   29/05/2013      
    12345       |   38170       |   Closed      |   31/05/2013      
    12345       |   38170       |   Closed      |   01/06/2013      
    12345       |   38170       |   Closed      |   03/06/2013      
    12345       |   38170       |   Reopened    |   04/06/2013  
    12345       |   38170       |   Reopened    |   04/06/2013      
    12345       |   38170       |   Reopened    |   07/06/2013      

而且，最好通过简单的SQL语句，而不是过程。

提前感谢。如有任何澄清，请评论

Indru

更新

当表B中的两行改变了状态，但在相同的日期，我在连接中得到不一致的值。

举个例子:

我有两张表，比如A &b:它们是相当大的表格，它们记录了一定的信息。表A每隔几天扫描一次数据，但缺少状态列如果发生了更改，那么表B将获得一个带有status列的新条目。我需要匹配这两个，并为表a中对应于b的每个条目派生一个状态。

表B

Id  | Num1  | Num2  | CreatedOn   | Status
100 | 12345 | 38170 | 28/05/2013  | New
150 | 12345 | 38170 | 28/05/2013  | Closed
200 | 12345 | 38170 | 31/05/2013  | Reopened
250 | 12345 | 38170 | 04/06/2013  | Closed

表

Id  | Num1  | Num2  | CreatedOn
55  | 12345 | 38170 | 28/05/2013
99  | 12345 | 38170 | 30/05/2013
145 | 12345 | 38170 | 31/05/2013
192 | 12345 | 38170 | 31/05/2013
223 | 12345 | 38170 | 04/06/2013

期望输出值

Id  | Num1  | Num2  | CreatedOn   | Status
55  | 11552 | 38170 | 28/05/2013  | New
99  | 11552 | 38170 | 30/05/2013  | Closed
145 | 11552 | 38170 | 31/05/2013  | Reopened
192 | 11552 | 38170 | 31/05/2013  | Reopened
223 | 11552 | 38170 | 04/06/2013  | Closed

相反，我得到:(使用杰弗里·坎普的回复)

Id  | Num1  | Num2  | CreatedOn   | Status
55  | 11552 | 38170 | 28/05/2013  | Closed
99  | 11552 | 38170 | 30/05/2013  | Closed
145 | 11552 | 38170 | 31/05/2013  | Reopened
192 | 11552 | 38170 | 31/05/2013  | Reopened
223 | 11552 | 38170 | 04/06/2013  | Closed

和这个:(使用回复2 - Mark Bannister)

Id  | Num1  | Num2  | CreatedOn   | Status
55  | 11552 | 38170 | 28/05/2013  | New
99  | 11552 | 38170 | 30/05/2013  | New
145 | 11552 | 38170 | 31/05/2013  | Reopened
192 | 11552 | 38170 | 31/05/2013  | Reopened
223 | 11552 | 38170 | 04/06/2013  | Closed

显然，记录Id-55既可以是New也可以是Closed，因为在表b中同一天(28/05/13)有两个条目。但逻辑是，它从New -> Closed ->重开有什么办法可以做到吗?

Answer 1

对于初学者来说很简单:

SELECT A.Num1
      ,A.Num2
      ,NVL(
         (SELECT DISTINCT
                 FIRST_VALUE(B.Status)
                 OVER (ORDER BY B.Date DESC)
          FROM   B
          WHERE  B.Num1 = A.Num1
          AND    B.Num2 = A.Num2
          AND    B.Date <= A.Date
         ),'New') AS Status
      ,A.Date
FROM A;

Answer 2

尝试:

select a.Num1, a.Num2, a."Date", b.Status
from TableA a
join (select TableB.*, 
             lead("Date",1) over (partition by Num1, Num2 
                                  order by "Date") NextDate
      from TableB) b
  on a.Num1 = b.Num1 and 
     a.Num2 = b.Num2 and 
     a."Date" >= b."Date" and 
     a."Date" < coalesce(b.NextDate,a."Date"+1)

SQLFiddle。