SqlServer性能优化索引

Posted 2020-08-25 石

导入表结构：

select * into ProductCategory from AdventureWorksDW2014.dbo.DimProductCategory
select * into Product from AdventureWorksDW2014.dbo.DimProduct

 

开启磁盘io：

set statistics io on
select EnglishProductName,StandardCost,Color,Size,Weight from Product
where size>\'M\'--0.189 io:251
set statistics io off

 

 

非聚簇索引：

创建的语句：
create nonclustered index nc_product_size on product(size)

 再次执行上面的查询代码（提高了三倍）：

set statistics io on
select EnglishProductName,StandardCost,Color,Size,Weight from Product
where size>\'M\'   --0.054 io:19
set statistics io off

 

建立覆盖索引：

create nonclustered index nc_product_size1 on product(size) include(EnglishProductName,
StandardCost,Color,Weight)

再次执行上述语句：

set statistics io on
select EnglishProductName,StandardCost,Color,Size,Weight from Product
where size>\'M\'   --0.003 io:2
set statistics io off

 数据库会自动选择索引：

 

没有创建索引的情况：

set statistics io on
select c.EnglishProductCategoryName,p.EnglishProductName,p.Color,p.Size
from product as p inner join ProductCategory as c on p.ProductSubcategoryKey=c.ProductCategoryKey
where c.ProductCategoryKey=1 --0.1928
set statistics io off

 创建索引：

create nonclustered index nc_productcategory_key on ProductCategory(ProductcategoryKey) include
(EnglishProductCategoryName)

 在次查询：

set statistics io on
select c.EnglishProductCategoryName,p.EnglishProductName,p.Color,p.Size
from product as p inner join ProductCategory as c on p.ProductSubcategoryKey=c.ProductCategoryKey
where c.ProductCategoryKey=1 --0.1928     io:c 2 p 251
set statistics io off

 IO情况：

由此可见 Product表影响比较严重 251

建立一个非聚簇索引：（做一个物理排序）

create nonclustered index nc_product_categorykey on product(productsubcategorykey) include
(englishproductname,color,size)

 执行语句：

set statistics io on
select c.EnglishProductCategoryName,p.EnglishProductName,p.Color,p.Size
from product as p inner join ProductCategory as c on p.ProductSubcategoryKey=c.ProductCategoryKey
where c.ProductCategoryKey=1      --4.29 od 1497 oh 783 c 155
set statistics io off

 

 

导入三张表：

select * into Customer from AdventureWorks2014.Sales.Customer
select * into OrderHeader from AdventureWorks2014.Sales.SalesOrderHeader
select * into OrderDetail from AdventureWorks2014.Sales.SalesOrderDetail

 实现一些业务：

set statistics io on 
select c.CustomerID,SUM(od.LineTotal) from OrderDetail as od inner join
 orderheader as oh  on od.SalesOrderID=oh.SalesOrderID  inner join customer as c
 on oh.CustomerID =c.CustomerID group by(c.CustomerID)   --4.29
 set statistics io off  

 

优化的第一步：

  1.查看sql语句写法是否有问题（进行改造）

set statistics io on
select oh.CustomerID,sum(od.LineTotal) from OrderDetail as od inner join
OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID) --3.77 od 1497 oh 783
set statistics io off

 

创建索引：

create nonclustered index nc_OrderDetail_SalesOrderID on OrderDetail(SalesOrderID) include
(linetotal)

 创建另外一个索引：针对group by 的列

create nonclustered index nc_OrderHeader_CustomerID on OrderHeader(CustomerID)

 

在次执行上述语句：

set statistics io on
select oh.CustomerID,sum(od.LineTotal) from OrderDetail as od inner join
OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID) --3.10 od 533 oh 783
set statistics io off

 

 

采用索引视图的方式：

create view v_Order_Total
as
select oh.CustomerID,sum(od.LineTotal) as 总额 from OrderDetail as od inner join
OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID)

 效果差不多：

set statistics io on
select * from v_Order_Total --3.10 od 533 oh 783
set statistics io off

 

修改：

alter view v_Order_Total
as
select oh.CustomerID as 客户ID, sum(od.LineTotal) as 总额 from OrderDetail as od inner join
OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID)

 对唯一列做聚集索引：

 create clustered index c_vordertotal_customerid on v_order_total(客户ID)

 直接运行报错：

解决方案：

在次执行：

alter view v_Order_Total
with schemabinding 
as
select oh.CustomerID as 客户ID, sum(od.LineTotal) as 总额 from OrderDetail as od inner join
OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID)

 报错：

解决方法：

alter view v_Order_Total
with schemabinding 
as
select oh.CustomerID as 客户ID, sum(od.LineTotal) as 总额 from dbo.OrderDetail as od inner join
dbo.OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID)

 在次创建：

 
 create clustered index c_vordertotal_customerid on v_order_total(客户ID)

 报错：

办法：

 create unique clustered index c_vordertotal_customerid on v_order_total(客户ID)

 报错：

方法：

alter view v_Order_Total
with schemabinding 
as
select oh.CustomerID as 客户ID, sum(od.LineTotal) as 总额,COUNT_BIG(*) as 计数 from dbo.OrderDetail as od inner join
dbo.OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID)

 执行创建索引：

 create unique clustered index c_vordertotal_customerid on v_order_total(客户ID)

 成功

set statistics io on
select * from v_Order_Total --0.09 io:92
set statistics io off

 执行计划：

 

 

 会自动进行值的更新，不用关心

对语句的访问会用到刚才的架构：

set statistics io on
select oh.CustomerID,sum(od.LineTotal) from OrderDetail as od inner join
OrderHeader as oh on od.SalesOrderID=oh.SalesOrderID group by(oh.CustomerID) --0.09 io:92
set statistics io off