DataBase 之 实用积累

(1)分页：（性能效果最好的是top/max 和Row_number()两种方法）

[1]写法1，not in/top

select top pagesize * from pagetest

where id not in 

    (select top (pagesize*(pageindex-1)) id 

     from pagetest order by id)

order by id

[2]写法2，not exists（引导的子句无结果集返回）

select top pagesize * from pagetest

where not exists

    (select 1 from 

        (select top (pagesize*(pageindex-1)) id from pagetest order by id) a  

     where a.id = pagetest.id)

order by id

[3]写法3，max/top

select top pagesize * from pagetest

where id>

    (select max(id) from 

        (select top (pagesize*(pageindex-1)) id 

         from pagetest order by id) a

     )

order by id

 

[4]写法4，row_number()

select top pagesize * from

    (select row_number()over(order by id)rownumber,* 

     from pagetest) a

where rownumber between startIndex and endIndex

[5]分页函数（先循环插入100000000条数据然后再分页）

create table  #FYTable (id int identity(1,1),title nvarchar(20),con nvarchar(20))

declare @title nvarchar(300),

        @con nvarchar(300)

declare @start int,@end int

set @start=0

set @end=1000000

set @title='Olive'

set @con='I love'

while @start<@end

begin

 insert into #FYTable select @title,@con

 set @start=@start+1

 end

select * from #FYTable

create procedure dbo.FenYe

@count int,@yeshu int

as

select top(@yeshu) * from #FYTable where id not in (select top(@count*@yeshu) id from #FYTable)

 

(2)排序：排序函数Rank()

select SName,Sage,SSex 

from (select SName,Sage,Ssex,Rank() over(order by sage) as rank 

      from Student) as a 

where a.rank > 1 and a.rank < 9;

(3)表行列转换：

数据表行列转换（case when then）

create table Factory( FID int, FDepartment nvarchar(20), FMeterial nvarchar(20),  FNumber int)

insert into Factory values(1,'工厂1','材料1',1);

insert into Factory values(1,'工厂2','材料1',1);

insert into Factory values(1,'工厂2','材料2',2);

insert into Factory values(1,'工厂2','材料3',1);

insert into Factory values(1,'工厂3','材料3',1);

insert into Factory values(1,'工厂2','材料1',2);

insert into Factory values(1,'工厂3','材料2',1);

select * from Factory

 

select  FDepartment,

SUM(case FMeterial when '材料1' then FNumber else 0 end) as 材料1,

SUM(case FMeterial when '材料2' then FNumber else 0 end) as 材料2,

SUM(case FMeterial when '材料3' then FNumber else 0 end) as 材料3

from Factory group by FDepartment

 

--行列转换函数Pivot

create table Sell

(

[Year] int , [Quarter] nvarchar(2) , Quantity int

)

insert into Sell values(2011,'Q1',12)

insert into Sell values(2011,'Q2',13)

insert into Sell values(2011,'Q2',14)

insert into Sell values(2011,'Q3',15)

insert into Sell values(2011,'Q2',12)

insert into Sell values(2011,'Q3',13)

insert into Sell values(2011,'Q4',15)

insert into Sell values(2011,'Q4',17)

insert into Sell values(2011,'Q4',11)

insert into Sell values(2012,'Q3',13)

insert into Sell values(2012,'Q4',15)

insert into Sell values(2012,'Q2',17)

insert into Sell values(2012,'Q4',11)

insert into Sell values(2012,'Q3',13)

insert into Sell values(2012,'Q1',15)

insert into Sell values(2012,'Q1',17)

insert into Sell values(2012,'Q4',11)

select * from Sell

 

select * from Sell Pivot(sum(Quantity) for [Quarter] in (Q1,Q2,Q3,Q4) )as p

select [Year], SUM(case when [Quarter]='Q1' then Quantity end) as Q1,

               SUM(case when [Quarter]='Q2' then Quantity end) as Q2,

               SUM(case when [Quarter]='Q3' then Quantity end) as Q3,

               SUM(case when [Quarter]='Q4' then Quantity end) as Q4

               from Sell group by [Year]

 

 

--case when then 用例

drop table T

go

create table T (TID int identity(1,1) not null  primary key,Quantity int not null,ProductID int not null,[Type] int not null)

go

insert into T select 2,12,1

insert into T select 6,12,2

insert into T select 3,10,2

insert into T select 8,10,1

insert into T select 2,12,1

select * from T

go

--统计每种商品有多少件

select ProductID,SUM (case when [Type]=1 then Quantity else -Quantity end) from T group by ProductID

 

--使用case when 为表的新列填充不同的数据

alter table T add Descriptions nvarchar(20) null

go

update T set Descriptions= case when [Type]=1 then '现存'

                                when [Type]=2 then '售出'

                                else '未知'

                                end

go               

 

--怎么把这样一个表

--year  month amount

--1991   1     1.1

--1991   2     1.2

--1991   3     1.3

--1991   4     1.4

--1992   1     2.1

--1992   2     2.2

--1992   3     2.3

--1992   4     2.4

--查成这样一个结果

--year  m1  m2  m3  m4

--1991  1.1   1.2   1.3   1.4

--1992  2.1   2.2   2.3   2.4

create table tb_Test1

([year] int,[month] int,amount float)

insert into tb_Test1

select 1991,1,1.1

union select 1991,2,1.2

union select 1991,3,1.3

union select 1991,4,1.4

union select 1992,1,1.1

union select 1992,2,1.2

union select 1992,3,1.3

union select 1992,4,1.4

select * from tb_Test1t

--根据月份的不同分为不同的字段，然后再按照年份分组

select [year],sum(case [MONTH] when 1 then amount else 0   end) as m1,sum(case [MONTH] when 2 then amount  else 0 end) as m2,sum(case [MONTH] when 3 then amount else 0 end) as m3,sum(case [MONTH] when 4 then amount else 0 end) as m4 from tb_Test1 group by [year]

 

--有两个表A和B，均有key和value两个字段，如果B的key在A中也有，就把B的value换为A中对应的value。这道题的SQL语句怎么写？

create table tb_A(keys int,value varchar(20))

create table tb_B(keys int,value varchar(20))

insert into tb_A values(1,'aa'),(2,'ab'),(3,'ac')

insert into tb_B values(1,'aa'),(2,'a'),(3,'a')

方法一--先将更新B中value的值更新为A中value的值，同时限定B中的keys=A.Keys 同时B.value<>A.value（不等于）

update tb_B set value=(select a.value from tb_A a where a.keys=tb_B.keys)

where tb_B.keys in(select tb_B.keys from tb_B,tb_A where tb_A.keys=tb_B.keys and tb_A.value<>tb_B.value)

select * from tb_B

select * from tb_A

delete from tb_A

delete from tb_B

方法二--先删除A，B一样的部分，然后再根据两表中Keys 的相等进行更新

update tb_B set tb_B.value=s.value from (select * from tb_A except select * from tb_B) as s where tb_B.keys=s.Keys

--原表:

--courseid  coursename  score

--1　　　  java 　　      70

--2 　　   oracle 　　    90

--3 　　   xml 　　      40

--4 　　   jsp 　　       30

--5 　　   servlet 　　    80

--为了便于阅读,查询此表后的结果显式如下(及格分数为60):

--courseid  coursename  score  mark

--1 　　     java 　　    70 　pass

--2  　　    oracle 　　  90 　pass

--3 　　     xml 　　    40 　fail

--4 　　     jsp 　　     30 　fail

--5 　　　  servlet 　　   80   pass

create table tb_B(id int identity(1,1),coursename varchar(30),score int )

insert into tb_B values('java',70),('oracle',90),('xml',40),('jsp',30),('servlet',80)

select * from tb_B

--case when 的使用

select id,coursename,score, case when score > 60 then 'pass' else 'fail' end as mark from tb_B                               

 

--原表:

--id proid proname

--1  1    M

--1  2    F

--2  1    N

--2  2    G

--3  1    B  

--3  2    A

--查询后的表:

--id  pro1  pro2

--1  M     F

--2  N     G

--3  B     A

create table C(id int,proid int,proname varchar(2))

insert into C

values(1,1,'M'),(1,2,'F'),(2,1,'N'),(2,2,'G'),(3,1,'B'),(3,2,'A')

select * from C

select id,(select proname

           from C c1

           where proid=1 and c.id=c1.id ) as pro1,

           (select proname from C c2 where proid=2 and c2.id=c.id ) as pro2

from C c group by id

--如下表a

--列    a1 a2

--记录  1  a

--      1  b

--      2  x

--      2  y

--      2  z

--用select能选成以下结果吗？

--1 ab

--2 xyz

create table D (a1 varchar(2),a2 varchar(2))

insert into  D values ('1','a'),('1','b'),('2','x'),('2','y'),('2','z')

Create table D1 (id varchar(2),value varchar(2))

Alter table D1 alter column value  varchar(20)

insert into D1 select Distinct(a1),'' from D

declare @id varchar(2),@value varchar(20)

declare LianJie cursor for select * from D

open LianJie

fetch next from LianJie into @id,@value

while @@FETCH_STATUS=0

begin

update D1 set value=value+@value where id=@id

fetch next from LianJie  into @id,@value

end

close LianJie

deallocate  LianJie

select * from D

select * from D1

 

——3. 表内容：

--2005-05-09 胜

--2005-05-09 胜

--2005-05-09 负

--2005-05-09 负

--2005-05-10 胜

--2005-05-10 负

--2005-05-10 负

--如果要生成下列结果, 该如何写sql语句?

--           胜   负

--2005-05-09  2    2

--2005-05-10  1    2

create table E (dt varchar(20),shengfu nvarchar(20))

insert into E values('2012-9-15','胜'),('2012-9-15','胜'),('2012-9-15','负'),('2012-9-15','负'),('2012-9-16','胜'),('2012-9-16','胜'),('2012-9-16','负')

select * from E

select  dt, SUM(case when shengfu='胜' then 1 else 0 end) as 胜,SUM(case when shengfu='负' then 1 else 0 end) as 负 from E group by dt

 

--表中有A B C三列,用SQL语句实现：当A列大于B列时选择A列否则选择B列，当B列大于C列时选择B列否则选择C列

create table E1 (A int ,B int,C int)

insert into E1 values(1,3,5),(6,4,2),(8,7,9)

select * from E1

select case when A>B then A else B end as AB,case when B>C then B else C end as BC from E1

 

table1

月份          部门业绩

一月份      01      10

一月份      02      10

一月份      03      5

二月份      02      8

二月份      04      9

三月份      03      8

table2

部门     部门名称

01      国内业务一部

02      国内业务二部

03      国内业务三部

04      国际业务部

table3 （result）

部门部门名称  一月份      二月份      三月份

  01  国内业务一部    10        null      null

  02   国内业务二部   10         8        null

  03   国内业务三部   null       5        8

  04   国际业务部   null      null      9

create table #F([month] nvarchar(20), bumen nvarchar(20), yeji int)

insert into #F values('一月份','01',10),('一月份','02',10),('一月份','03',5),('二月份','02',8),('二月份','04',9),('三月份','03',8)

create table #F1(bumen nvarchar(20),bumenmingcheng nvarchar(20))

insert into #F1 values('01','国内业务一部'),('02','国内业务二部'),('03','国内业务三部'),('04','国际业务部')

select * from #F1

select f.bumen ,f1.bumenmingcheng,sum(case when f.[month]='一月份' then f.yeji else null end) as 一月份,

sum(case when f.[month]='二月份' then f.yeji else null end) as 二月份,

sum(case when f.[month]='三月份' then f.yeji else null end) as 三月份 from #F f,#F1 f1 where f.bumen=f1.bumen group by f.bumen,f1.bumenmingcheng

 

drop table #Str

create table #Str(id int identity(1,1),str11 varchar(20), [type] int )

insert into #Str values('how',1),('are',1),('you',1),('fine',2),('thank',2),('you',2)

select * from #Str

create table #Str1(ss nvarchar(100),id int )

insert into #Str1 select Distinct '',([type]) from #Str

select * from #Str1

declare @id1 int,@str1 nvarchar(100)

declare StrCursor cursor for select str11,[type] from #Str

open StrCursor

fetch next from StrCursor into @str1,@id1

while @@FETCH_STATUS=0

begin

 update #Str1 set ss=ss+@str1 whereid=@id1

 fetch next from StrCursor into @str1,@id1

 

 end

 close StrCursor

 deallocate StrCursor

 select * from #Str1

 

 

 --Post

 create Table Post(id int identity(1,1) primary key,title nvarchar(200),content text,CreateDate nvarchar(30))

 --Comments

 create Table Comments(Cid int identity(1,1) primary key,id int,content text,CreateDate nvarchar(30))

insert into Post select '我从远方赶来','我为她而狂野，我为你来看我不顾一切',cast(GETDATE() as nvarchar)

insert into Comments values

(1,'不虚此行呀',CAST(getdate() as nvarchar)),

(1,'这一个不能停留太久的世界',cast(getdate() as nvarchar)),

(1,'无可取代',cast(getdate() as nvarchar))

select * from Comments

create table #Ts (id int,com nvarchar(max))

insert into #Ts select Distinct(id),'' from Comments

declare @ii int,@com nvarchar(max)

declare TsCursor cursor for select id ,content from Comments

open TsCursor

fetch next from TsCursor into @ii,@com

while @@FETCH_STATUS=0

begin

update #Ts set com=com+@com whereid=@ii

fetch next from TsCursor into @ii,@com

end

close TsCursor

deallocate TsCursor