Question

我有一个看起来像这样的SQL Server表（ID是PK标识字段）


  ID Date                  Condition     Call_Count    
  1  2015-03-02 14:05:05   Terminated     0    
  2  2015-03-02 14:05:10   Transferred    0    
  3  2015-03-02 14:06:02   Transferred    0    
  4  2015-03-02 14:07:00   Transferred    0    
  5  2015-03-02 14:07:50   Terminated     0    
  6  2015-03-02 14:07:60   Transferred    0    
  7  2015-03-02 14:08:55   Terminated     0    
  8  2015-03-02 14:08:56   Terminated     0

我试过以下内容：

DECLARE @LoopId INT = 1, 
    @Call_Count INT = 1, 
    @callend INT,
    @maxId INT,
    @LoopCounter INT

Select @MaxId = max(ID) from Table1     

WHILE(@LoopId <= @MaxId)
BEGIN
   SELECT @callend = charindex('Terminated', Reason)
     from Table1 where ID = @LoopId

   set @Call_Count = @Call_Count + case when @callend > 0 then 1 else 0 end

   IF @LoopId > 1 
   begin
     update Table1
        set Call_Counter = @Call_Count
      WHERE Id = @LoopID
   end

   SET @LoopId  = @LoopId  + 1        
END

我得到的结果是：


  ID Date                  Condition     Call_Count    
  1   2015-03-02 14:05:05   Terminated     1  
  2   2015-03-02 14:05:05   Terminated     2
  3   2015-03-02 14:05:05   Terminated     3
  4   2015-03-02 14:05:05   Terminated     4  
  5   2015-03-02 14:05:10   Transferred    4    
  6   2015-03-02 14:06:02   Transferred    4    
  7   2015-03-02 14:07:00   Transferred    4    
  8   2015-03-02 14:07:50   Terminated     5    
  9   2015-03-02 14:07:60   Transferred    5    
  10  2015-03-02 14:08:55   Terminated     6    
  11  2015-03-02 14:08:56   Terminated     6

我需要生成一个会产生此结果的ID字段，它需要在最后一个“Terminated”之后更改数字，以便将一个“call”放入一个组中。


  ID Date                  Condition     Call_Count    
  1   2015-03-02 14:05:05   Terminated     1  
  2   2015-03-02 14:05:05   Terminated     2
  3   2015-03-02 14:05:05   Terminated     3
  4   2015-03-02 14:05:05   Terminated     4  
  5   2015-03-02 14:05:10   Transferred    5    
  6   2015-03-02 14:06:02   Transferred    5    
  7   2015-03-02 14:07:00   Transferred    5    
  8   2015-03-02 14:07:50   Terminated     5    
  9   2015-03-02 14:07:60   Transferred    6    
  10  2015-03-02 14:08:55   Terminated     6    
  11  2015-03-02 14:08:56   Terminated     7

Answer 1

应该适用于任何版本的sql server的简单解决方案是使用带有~~的内部联接到相关派生表的更新~~ 相关的子查询。

创建并填充样本表（请在将来的问题中将此步骤保存起来 - 如果您自己这样做，我会在第一次尝试中发布正确且经过测试的解决方案）

CREATE TABLE Table1
(
    ID int identity(1,1),
    [Date] datetime,
    Condition nvarchar(20),
    Call_Counter int
)
INSERT INTO Table1 ([Date], Condition) VALUES 
('2015-03-02 14:05:05', 'Terminated'),
('2015-03-02 14:05:10', 'Transferred'),
('2015-03-02 14:06:02', 'Transferred'),
('2015-03-02 14:07:00', 'Transferred'),
('2015-03-02 14:07:50', 'Terminated'),
('2015-03-02 14:07:59', 'Transferred'),
('2015-03-02 14:08:55', 'Terminated'),
('2015-03-02 14:08:56', 'Terminated')

更新声明：

UPDATE t1
SET Call_Counter = 
    (
        SELECT COUNT(*) As CallCounter
        FROM Table1 t2
        WHERE Condition = 'Terminated'
        AND t2.ID <= t1.ID
    ) 
FROM Table1 t1

See live demo on rextester.

第一次尝试（错误） <击>

<击>

UPDATE t1
SET Call_Counter = CallCounter
FROM Table1 t1
INNER JOIN 
(
    SELECT COUNT(*) As CallCounter
    FROM Table1 t2
    WHERE Condition = 'Terminated'
    AND ID <= t1.ID
) t3

<击>

Answer 2

类似的解决方案让Zohar。但我宁愿先计算所有的价值。还可以利用窗口功能。

<强> SQL DEMO

WITH CTE as (
    SELECT ID , 
           COUNT(CASE WHEN [Condition] = 'Terminated' THEN 1 END) 
           OVER (ORDER BY [ID]) as rn
    FROM Table1
)      
UPDATE A
SET [Call_Count] =  CASE WHEN [Condition] = 'Terminated' 
                         THEN B.rn
                         ELSE B.rn + 1
                    END
FROM Table1 A
JOIN CTE B
  ON A.ID = B.ID;

将数字添加到SQL Server表中，该表在条件上递增

2 个答案: