SQL Server2008中刪除重復(fù)記錄的方法分享

更新時(shí)間：2011年10月31日 23:30:37 作者：

在Database中可能由于某種原因如用戶輸入,導(dǎo)入數(shù)據(jù)失敗等導(dǎo)致了重復(fù)記錄. 如果你沒有用主鍵,約束,或來其它機(jī)制實(shí)現(xiàn)數(shù)據(jù)完整性,那最后總是重復(fù)記錄在你的數(shù)據(jù)庫中.

現(xiàn)在讓我們來看在SQL SERVER 2008中如何刪除這些記錄, 首先,可以模擬造一些簡單重復(fù)記錄:

復(fù)制代碼代碼如下:

 
Create Table dbo.Employee ( 
[Id] int Primary KEY , 
[Name] varchar(50), 
[Age] int, 
[Sex] bit default 1 
) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(1,'James',25,default) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(2,'James',25,default) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(3,'James',25,default) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(4,'Lisa',24,0) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(5,'Lisa',24,0) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(6,'Lisa',24,0) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(7,'Mirsa',23,0) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(8,'Mirsa',23,0) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(9,'Mirsa',23,0) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(10,'John',26,default) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(11,'Abraham',28,default) 
Insert Into Employee ([Id] , [Name] , [Age] , [Sex] ) Values(12,'Lincoln',30,default) 

OK,首先我們使用最常見的方法:

Delete From Employee Where Name in (select NameFrom Employee Group By Name Having Count(Name)>1);
接著使用RowNumber():

Delete T From( Select Row_Number() Over(Partition By [Name] Order By (SELECT 0)) As RowNumber,* From Employee) TWhere T.RowNumber > 1;

還可以使用CTE (Common Table Expressions):

復(fù)制代碼代碼如下:

 
With Dups as 
( 
select ROW_NUMBER() Over(Partition by [Name] Order by (SELECT 0)) as rn 
FROM Employee 
) 
Delete From Dups 
Where rn>1; 

再加上RANK()的CTE:

復(fù)制代碼代碼如下:

 
WITH Dups As 
( 
Select [ID],[Name],[Age],[Sex] 
, ROW_NUMBER() OVER(Partition By [Name] Order By (SELECT 0)) AS rn 
,RANK() OVER(Partition By [Name] Order By (SELECT 0)) AS rnk 
FROM Employee 
) 
DELETE FROM Dups 
WHERE rn<>rnk; 

下面是這四個(gè)T-SQL查詢的執(zhí)行計(jì)劃:

你可以看到?jīng)]有用CTE的方法開銷最大, 主要是在Table Spool, 這里開銷了44%, Table Spool 是一個(gè)物理運(yùn)算符。

Table Spool 運(yùn)算符掃描輸入，并將各行的一個(gè)副本放入隱藏的假脫機(jī)表中，此表存儲(chǔ)在 tempdb 數(shù)據(jù)庫中并且僅在查詢的生存期內(nèi)存在。如果重繞該運(yùn)算符（例如通過 Nested Loops 運(yùn)算符重繞），但不需要任何重新綁定，則將使用假脫機(jī)數(shù)據(jù)，而不用重新掃描輸入。
注意上面的方法只是在重復(fù)記錄比較少的情況下, 如果重復(fù)記錄多. DELETE將會(huì)非常慢, 最好的方法是復(fù)制目標(biāo)數(shù)據(jù)到另一個(gè)新表,刪除原來的表,重命名新表為原來的表. 或用臨時(shí)表, 這樣還可以減少數(shù)據(jù)庫事務(wù)日志. 看下面的T-SQL:

復(fù)制代碼代碼如下:

 
WITH Dups As 
( 
Select [ID],[Name],[Age],[Sex] 
, ROW_NUMBER() OVER(Partition By [ID] Order By (SELECT 0)) AS rn 
FROM Employee 
) 
Select [ID],[Name],[Age],[Sex] 
INTO dbo.EmployeeDupsTmp 
FROM Dups 
WHERE rn=1 
DROP TABLE dbo.Employee; 
EXEC sp_rename 'dbo.EmployeeDupsTmp','Employee' 

希望這篇POST對(duì)您開發(fā)有幫助.作者：Petter Liu

您可能感興趣的文章: