Sql server 第一次出现重复的SQL更新

我有一个数据库，它是由一个实用程序填充的，该实用程序不能很好地处理同时请求

在这些表中插入行的逻辑存在缺陷，从而允许在派生表中重复。我的目标是修复这些副本

出现了两种类型的重复项：

钥匙是重复的，但我们可以用号码告诉他们 分开

其中键和数字都是重复的，在本例中为I 要更新第一个事件

情况是这样的

/** Sequence of events **/


/** First Set (Correct) **/

Insert into Base values(100,'Base1');
    /** Find the biggest IKey in Base Table **/
    /** Set that as IKey in Derived Table (IKey = 1)**/
Insert into Derived values(1,100,'Derived1');

Insert into Base values(101,'Base2');
    /** Find the biggest IKey in Base Table **/
    /** Set that as IKey in Derived Table (IKey = 2)**/
Insert into Derived values(2,101,'Derived2');



/** Second Set ( duplicate IKey )**/

Insert into Base values(200,'Base3');
    /** Find the biggest IKey in Base Table **/
    /** We would have gotten 3, but there was another insert before this query could be run **/
Insert into Base values(201,'Base4');
    /** Because of the insert, we got a 4 instead of a 3 **/


    /** Set that as IKey in Derived Table (IKey = 4)**/
Insert into Derived values(4,200,'Derived3');
    /** Find the biggest IKey in Base Table **/
    /** Set that as IKey in Derived Table (IKey = 4)**/
Insert into Derived values(4,201,'Derived4');




/** Third Set ( duplicate IKey and Number )**/

Insert into Base values(300,'Base5');
    /** Find the biggest IKey in Base Table **/
    /** We would have gotten 5, but there was another insert before this query could be run **/
    /** This even had the "Number" column same as previous insert **/ 

Insert into Base values(300,'Base6');
    /** Because of the insert, we got a 6 instead of a 5 **/
    /** Set that as IKey in Derived Table (IKey = 6)**/

Insert into Derived values(6,300,'Derived5');
    /** Find the biggest IKey in Base Table **/
    /** Set that as IKey in Derived Table (IKey = 6)**/
Insert into Derived values(6,300,'Derived6');

结果如下表所示：

基地

衍生

IKey Number Derived
1   100     Derived1
2   101     Derived2
4   200     Derived3
4   201     Derived4
6   300     Derived5
6   300     Derived6

我已经找到了这个案件的解决办法

我想为#2找到一个解决方案

这里是一个工作区，用来试验这个问题。

---

如果您使用

行数（）分区方式（）

函数，那么您将获得“组”（相同标识符）的编号；然后，您可以使用每个组的第2行更新第1行。如果您使用

row#u number（）分区by（）

函数，则可以获得“组”的编号（相同标识符）；然后，您可以使用每个组的行2更新行1，每个组的行可能重复

IKey Number Derived
1   100     Derived1
2   101     Derived2
4   200     Derived3
4   201     Derived4
6   300     Derived5
6   300     Derived6

UPDATE Derived
Set Derived.IKey = Derived.IKey - 1

FROM    Derived 
        INNER JOIN
        Base
        ON
            Derived.IKey - 1 = Base.IKey
            AND 
            Derived.Number = Base.Number

WHERE   Derived.IKey
IN
      (
              /** Query to get all records which have duplicates in a table but the previous key is not used**/
              SELECT
                Derived.IKey
              FROM
                Derived
              WHERE
                Derived.IKey - 1 NOT IN (SELECT IKey FROM Derived)
              GROUP BY 
                Derived.IKey
              HAVING 
                COUNT(Derived.IKey) > 1
       )