查询及删除重复记录的方法

水深无声 2021-11-11 23:30 328阅读 0赞

![None.gif][] (一)  
![None.gif][] **1** 、查找表中多余的重复记录，重复记录是根据单个字段（peopleId）来判断  
![None.gif][] select   \*   from  people  
![None.gif][] where  peopleId  in  ( select   peopleId   from   people   group    by   peopleId   having    count (peopleId)  >   **1** )  
![None.gif][]  
![None.gif][] **2** 、删除表中多余的重复记录，重复记录是根据单个字段（peopleId）来判断，只留有rowid最小的记录  
![None.gif][] delete   from  people   
![None.gif][] where  peopleId   in  ( select   peopleId   from  people   group    by   peopleId    having    count (peopleId)  >   **1** )  
![None.gif][] and  rowid  not   in  ( select   min (rowid)  from   people   group   by  peopleId   having   count (peopleId ) > **1** )  
![None.gif][]  
![None.gif][] **3** 、查找表中多余的重复记录（多个字段）   
![None.gif][] select   \*   from  vitae a  
![None.gif][] where  (a.peopleId,a.seq)  in   ( select  peopleId,seq  from  vitae  group   by  peopleId,seq   having   count ( \* )  >   **1** )  
![None.gif][]  
![None.gif][] **4** 、删除表中多余的重复记录（多个字段），只留有rowid最小的记录  
![None.gif][] delete   from  vitae a  
![None.gif][] where  (a.peopleId,a.seq)  in   ( select  peopleId,seq  from  vitae  group   by  peopleId,seq  having   count ( \* )  >   **1** )  
![None.gif][] and  rowid  not   in  ( select   min (rowid)  from  vitae  group   by  peopleId,seq  having   count ( \* ) > **1** )  
![None.gif][]  
![None.gif][]  
![None.gif][] **5** 、查找表中多余的重复记录（多个字段），不包含rowid最小的记录  
![None.gif][] select   \*   from  vitae a  
![None.gif][] where  (a.peopleId,a.seq)  in   ( select  peopleId,seq  from  vitae  group   by  peopleId,seq  having   count ( \* )  >   **1** )  
![None.gif][] and  rowid  not   in  ( select   min (rowid)  from  vitae  group   by  peopleId,seq  having   count ( \* ) > **1** )  
![None.gif][]  
![None.gif][](二)  
![None.gif][]比方说  
![None.gif][]在A表中存在一个字段“name”，  
![None.gif][]而且不同记录之间的“name”值有可能会相同，  
![None.gif][]现在就是需要查询出在该表中的各记录之间，“name”值存在重复的项；  
![None.gif][] Select  Name, Count ( \* )  From  A  Group   By  Name  Having   Count ( \* )  >   **1**   
![None.gif][]  
![None.gif][]如果还查性别也相同大则如下:  
![None.gif][] Select  Name,sex, Count ( \* )  From  A  Group   By  Name,sex  Having   Count ( \* )  >   **1**   
![None.gif][]  
![None.gif][]  
![None.gif][](三)  
![None.gif][]方法一  
![None.gif][]  
![None.gif][] declare   @max   **integer** , @id   **integer**   
![None.gif][]  
![None.gif][] declare  cur\_rows  cursor  local  for   select  主字段, count ( \* )  from  表名  group   by  主字段  having   count ( \* )  > ；  **1**   
![None.gif][]  
![None.gif][] open  cur\_rows  
![None.gif][]  
![None.gif][] fetch  cur\_rows  into   @id , @max   
![None.gif][]  
![None.gif][] while   **@@fetch\_status** = **0**   
![None.gif][]  
![None.gif][] begin   
![None.gif][]  
![None.gif][] select   @max   =   @max   \- **1**   
![None.gif][]  
![None.gif][] set   rowcount   @max   
![None.gif][]  
![None.gif][] delete   from  表名  where  主字段  =   @id   
![None.gif][]  
![None.gif][] fetch  cur\_rows  into   @id , @max   
![None.gif][]  
![None.gif][] end   
![None.gif][]  
![None.gif][] close  cur\_rows  
![None.gif][]  
![None.gif][] set   rowcount   **0**   
![None.gif][]  
![None.gif][]　　方法二  
![None.gif][]  
![None.gif][]　　有两个意义上的重复记录，一是完全重复的记录，也即所有字段均重复的记录，二是部分关键字段重复的记录，比如Name字段重复，而其他字段不一定重复或都重复可以忽略。  
![None.gif][]  
![None.gif][]　　 **1** 、对于第一种重复，比较容易解决，使用  
![None.gif][]  
![None.gif][] select   distinct   \*   from  tableName  
![None.gif][]  
![None.gif][]　　就可以得到无重复记录的结果集。  
![None.gif][]  
![None.gif][]　　如果该表需要删除重复的记录（重复记录保留1条），可以按以下方法删除  
![None.gif][]  
![None.gif][] select   distinct   \*   into  \#Tmp  from  tableName  
![None.gif][]  
![None.gif][] drop   table  tableName  
![None.gif][]  
![None.gif][] select   \*   into  tableName  from  \#Tmp  
![None.gif][]  
![None.gif][] drop   table  \#Tmp  
![None.gif][]  
![None.gif][]　　发生这种重复的原因是表设计不周产生的，增加唯一索引列即可解决。  
![None.gif][]  
![None.gif][]　　 **2** 、这类重复问题通常要求保留重复记录中的第一条记录，操作方法如下  
![None.gif][]  
![None.gif][]　　假设有重复的字段为Name,Address，要求得到这两个字段唯一的结果集  
![None.gif][]  
![None.gif][] select   identity ( **int** , **1** , **1** )  as  autoID,  \*   into  \#Tmp  from  tableName  
![None.gif][]  
![None.gif][] select   min (autoID)  as  autoID  into  \#Tmp2  from  \#Tmp  group   by  Name,autoID  
![None.gif][]  
![None.gif][] select   \*   from  \#Tmp  where  autoID  in ( select  autoID  from  \#tmp2)  
![None.gif][]  
![None.gif][]　　最后一个select即得到了Name，Address不重复的结果集（但多了一个autoID字段，实际写时可以写在select子句中省去此列）  
![None.gif][]  
![None.gif][](四)  
![None.gif][]查询重复  
![None.gif][]  
![None.gif][] select   \*   from  tablename  where  id  in  (  
![None.gif][]  
![None.gif][] select  id  from  tablename   
![None.gif][]  
![None.gif][] group   by  id   
![None.gif][]  
![None.gif][] having   count (id)  >   **1**   
![None.gif][]  
![None.gif][])

转载于:https://www.cnblogs.com/Dicky/archive/2006/08/13/475592.html

[None.gif]: