sql分组后二次汇总(处理表重复记录查询和删除)的实现方法
--处理表重复记录(查询和删除)
/******************************************************************************************************************************************************
1、Num、Name相同的重复值记录,没有大小关系只保留一条
2、Name相同,ID有大小关系时,保留大或小其中一个记录
整理人：中国风(Roy)
日期:2008.06.06
******************************************************************************************************************************************************/
--1、用于查询重复处理记录(如果列没有大小关系时2000用生成自增列和临时表处理,SQL2005用row_number函数处理)
-->-->(Roy)生成測試數據
ifnotobject_id('Tempdb..#T')isnull
droptable#T
Go
Createtable#T([ID]int,[Name]nvarchar(1),[Memo]nvarchar(2))
Insert#T
select1,N'A',N'A1'unionall
select2,N'A',N'A2'unionall
select3,N'A',N'A3'unionall
select4,N'B',N'B1'unionall
select5,N'B',N'B2'
Go
--I、Name相同ID最小的记录(推荐用1,2,3),方法3在SQl05时，效率高于1、2
方法1:
Select*from#Tawherenotexists(select1from#TwhereName=a.NameandID<a.ID)
方法2:
selecta.*from#Tajoin(selectmin(ID)ID,Namefrom#TgroupbyName)bona.Name=b.Nameanda.ID=b.ID
方法3:
select*from#TawhereID=(selectmin(ID)from#TwhereName=a.Name)
方法4:
selecta.*from#Tajoin#Tbona.Name=b.Nameanda.ID>=b.IDgroupbya.ID,a.Name,a.Memohavingcount(1)=1
方法5:
select*from#TagroupbyID,Name,MemohavingID=(selectmin(ID)from#TwhereName=a.Name)
方法6:
select*from#Tawhere(selectcount(1)from#TwhereName=a.NameandID<a.ID)=0
方法7:
select*from#TawhereID=(selecttop1IDfrom#TwhereName=a.nameorderbyID)
方法8:
select*from#TawhereID!>all(selectIDfrom#TwhereName=a.Name)
方法9(注:ID为唯一时可用):
select*from#TawhereIDin(selectmin(ID)from#TgroupbyName)
--SQL2005:
方法10:
selectID,Name,Memofrom(select*,min(ID)over(partitionbyName)asMinIDfrom#Ta)TwhereID=MinID
方法11:
selectID,Name,Memofrom(select*,row_number()over(partitionbyNameorderbyID)asMinIDfrom#Ta)TwhereMinID=1
生成结果:
/*
ID NameMemo
-------------------
1 A A1
4 B B1
(2行受影响)
*/
--II、Name相同ID最大的记录,与min相反:
方法1:
Select*from#Tawherenotexists(select1from#TwhereName=a.NameandID>a.ID)
方法2:
selecta.*from#Tajoin(selectmax(ID)ID,Namefrom#TgroupbyName)bona.Name=b.Nameanda.ID=b.IDorderbyID
方法3:
select*from#TawhereID=(selectmax(ID)from#TwhereName=a.Name)orderbyID
方法4:
selecta.*from#Tajoin#Tbona.Name=b.Nameanda.ID<=b.IDgroupbya.ID,a.Name,a.Memohavingcount(1)=1
方法5:
select*from#TagroupbyID,Name,MemohavingID=(selectmax(ID)from#TwhereName=a.Name)
方法6:
select*from#Tawhere(selectcount(1)from#TwhereName=a.NameandID>a.ID)=0
方法7:
select*from#TawhereID=(selecttop1IDfrom#TwhereName=a.nameorderbyIDdesc)
方法8:
select*from#TawhereID!<all(selectIDfrom#TwhereName=a.Name)
方法9(注:ID为唯一时可用):
select*from#TawhereIDin(selectmax(ID)from#TgroupbyName)
--SQL2005:
方法10:
selectID,Name,Memofrom(select*,max(ID)over(partitionbyName)asMinIDfrom#Ta)TwhereID=MinID
方法11:
selectID,Name,Memofrom(select*,row_number()over(partitionbyNameorderbyIDdesc)asMinIDfrom#Ta)TwhereMinID=1
生成结果2:
/*
ID NameMemo
-------------------
3 A A3
5 B B2
(2行受影响)
*/
--2、删除重复记录有大小关系时,保留大或小其中一个记录
-->-->(Roy)生成測試數據
ifnotobject_id('Tempdb..#T')isnull
droptable#T
Go
Createtable#T([ID]int,[Name]nvarchar(1),[Memo]nvarchar(2))
Insert#T
select1,N'A',N'A1'unionall
select2,N'A',N'A2'unionall
select3,N'A',N'A3'unionall
select4,N'B',N'B1'unionall
select5,N'B',N'B2'
Go
--I、Name相同ID最小的记录(推荐用1,2,3),保留最小一条
方法1:
deleteafrom#Tawhere exists(select1from#TwhereName=a.NameandID<a.ID)
方法2:
deletea from#Taleftjoin(selectmin(ID)ID,Namefrom#TgroupbyName)bona.Name=b.Nameanda.ID=b.IDwhereb.Idisnull
方法3:
deleteafrom#TawhereIDnotin(selectmin(ID)from#TwhereName=a.Name)
方法4(注:ID为唯一时可用):
deleteafrom#TawhereIDnotin(selectmin(ID)from#TgroupbyName)
方法5:
deleteafrom#Tawhere(selectcount(1)from#TwhereName=a.NameandID<a.ID)>0
方法6:
deleteafrom#TawhereID<>(selecttop1IDfrom#TwhereName=a.nameorderbyID)
方法7:
deleteafrom#TawhereID>any(selectIDfrom#TwhereName=a.Name)
select*from#T
生成结果:
/*
ID NameMemo
-------------------
1 A A1
4 B B1
(2行受影响)
*/
--II、Name相同ID保留最大的一条记录:
方法1:
deleteafrom#Tawhere exists(select1from#TwhereName=a.NameandID>a.ID)
方法2:
deletea from#Taleftjoin(selectmax(ID)ID,Namefrom#TgroupbyName)bona.Name=b.Nameanda.ID=b.IDwhereb.Idisnull
方法3:
deleteafrom#TawhereIDnotin(selectmax(ID)from#TwhereName=a.Name)
方法4(注:ID为唯一时可用):
deleteafrom#TawhereIDnotin(selectmax(ID)from#TgroupbyName)
方法5:
deleteafrom#Tawhere(selectcount(1)from#TwhereName=a.NameandID>a.ID)>0
方法6:
deleteafrom#TawhereID<>(selecttop1IDfrom#TwhereName=a.nameorderbyIDdesc)
方法7:
deleteafrom#TawhereID<any(selectIDfrom#TwhereName=a.Name)
select*from#T
/*
ID NameMemo
-------------------
3 A A3
5 B B2
(2行受影响)
*/
--3、删除重复记录没有大小关系时，处理重复值
-->-->(Roy)生成測試數據
ifnotobject_id('Tempdb..#T')isnull
droptable#T
Go
Createtable#T([Num]int,[Name]nvarchar(1))
Insert#T
select1,N'A'unionall
select1,N'A'unionall
select1,N'A'unionall
select2,N'B'unionall
select2,N'B'
Go
方法1:
ifobject_id('Tempdb..#')isnotnull
droptable#
Selectdistinct*into#from#T--排除重复记录结果集生成临时表#
truncatetable#T--清空表
insert#Tselect*from# --把临时表#插入到表#T中
--查看结果
select*from#T
/*
Num Name
---------------
1 A
2 B
(2行受影响)
*/
--重新执行测试数据后用方法2
方法2:
altertable#TaddIDintidentity--新增标识列
go
deleteafrom #Tawhere exists(select1from#TwhereNum=a.NumandName=a.NameandID>a.ID)--只保留一条记录
go
altertable#TdropcolumnID--删除标识列
--查看结果
select*from#T
/*
Num Name
---------------
1 A
2 B
(2行受影响)
*/
--重新执行测试数据后用方法3
方法3:
declareRoy_Cursorcursorlocalfor
selectcount(1)-1,Num,Namefrom#TgroupbyNum,Namehavingcount(1)>1
declare@conint,@Numint,@Namenvarchar(1)
openRoy_Cursor
fetchnextfromRoy_Cursorinto@con,@Num,@Name
while@@Fetch_status=0
begin
setrowcount@con;
delete#TwhereNum=@NumandName=@Name
setrowcount0;
fetchnextfromRoy_Cursorinto@con,@Num,@Name
end
closeRoy_Cursor
deallocateRoy_Cursor
--查看结果
select*from#T
/*
Num Name
---------------
1 A
2 B
(2行受影响)