第十二章SQLServer统计信息(2)非索引键上统计信息的影
前言: 索引 对性能方面总是扮演着一个重要的角色,实际上,查询优化器首先检查谓词上的 统计 信息 ,然后才决定用什么 索引 。一般情况下,默认会在创建 索引 时, 索引 列上均创建 统计 信息 。但是不代表在非 索引 键上的 统计 信息 对性能没有用。 如果
前言:
索引对性能方面总是扮演着一个重要的角色,实际上,查询优化器首先检查谓词上的统计信息,然后才决定用什么索引。一般情况下,默认会在创建索引时,索引列上均创建统计信息。但是不代表在非索引键上的统计信息对性能没有用。
如果表上的所有列都有索引,那么将会是数据库负担不起,同时也不是一个好想法,包括谓词中用到的所有列加索引同样也不是好方法。因为索引会带来负载。因为需要空间存放索引,且每个DML语句都会需要更新索引。
一般来说,建议在where或者ON子句中出现的列上添加索引,但是由于某些情况,很难在所有的谓词上都创建索引,此时创建统计信息会是一个最起码的改进。如果Auto_Create_Statistics为ON,那么优化器会帮你做这一步。
准备工作:
默认情况下,Auto_Create_Statistics在数据库级别是设为ON的,但是为了下面需要这里先改成OFF:
ALTER DATABASE AdventureWorks2012 SET AUTO_CREATE_STATISTICS OFF GO ALTER DATABASE AdventureWorks2012 SET AUTO_UPDATE_STATISTICS OFF GO
然后创建一个新表用于本文使用:
SELECT * INTO SalesOrdDemo FROM Sales.SalesOrderHeader GO
步骤:
1、 对于新表,现在是没有统计信息在上面的,可以使下面语句来验证:
SELECT object_id , OBJECT_NAME(object_id) AS TableName , name AS StatisticsName , auto_created FROM sys.stats WHERE object_id = OBJECT_ID('SalesOrdDemo') ORDER BY object_id DESC GO
因为没有统计信息,所以这个查询是没有数据的。
2、 现在在新表上创建一个聚集索引:
CREATE CLUSTERED INDEX idx_SalesOrdDemo_SalesOrderID ON SalesOrdDemo(SalesOrderID) GO
3、 再次运行步骤一的脚本,可以看到已经有了数据,现在来执行下面的语句,并开启执行计划:
SELECT s.salesorderid , so.SalesOrderDetailID FROM salesordDemo AS s INNER JOIN Sales.SalesOrderDetail AS so ON s.salesorderid = so.SalesOrderID WHERE s.duedate = '2005-09-19 00:00:00.000'
4、 下面截图是步骤3中的执行计划,关注一下SalesOrdDemo表上有聚集索引扫描,这是合理的,因为没有WHERE子句在使用SalesOrderID列。而SalesOrderDetails表有非聚集索引扫描。还可以看到实际行数和估计行数有很大差异。
5、 现在是时候在新表的DueDate上创建统计信息,因为在查询中这个列并不包含在索引里面。
CREATE STATISTICS st_SaledOrdDemo_DueDate ON SalesOrdDemo(DueDate) GO
6、 再次执行步骤3的脚本,不需要任何改动:
SELECT s.salesorderid , so.SalesOrderDetailID FROM salesordDemo AS s INNER JOIN Sales.SalesOrderDetail AS so ON s.salesorderid = so.SalesOrderID WHERE s.duedate = '2005-09-19 00:00:00.000'
7、 对比上面的执行计划,此时在SalesOrderDetails表上已经从非聚集索引扫描变成了聚集索引查找,且开销只有2%,更总要的是实际行数和预估行数相差无几:
分析:
如果优化器可以获得谓词上列的统计信息,那么相会知道将要返回的行数,并且帮助优化器选择最佳的执行方式。

Hot AI Tools

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Undress AI Tool
Undress images for free

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics











The import steps are as follows: Copy the MDF file to SQL Server's data directory (usually C:\Program Files\Microsoft SQL Server\MSSQL\DATA). In SQL Server Management Studio (SSMS), open the database and select Attach. Click the Add button and select the MDF file. Confirm the database name and click the OK button.

For objects with the same name that already exist in the SQL Server database, the following steps need to be taken: Confirm the object type (table, view, stored procedure). IF NOT EXISTS can be used to skip creation if the object is empty. If the object has data, use a different name or modify the structure. Use DROP to delete existing objects (use caution, backup recommended). Check for schema changes to make sure there are no references to deleted or renamed objects.

To view the SQL Server port number: Open SSMS and connect to the server. Find the server name in Object Explorer, right-click it and select Properties. In the Connection tab, view the TCP Port field.

If you accidentally delete a SQL Server database, you can take the following steps to recover: stop database activity; back up log files; check database logs; recovery options: restore from backup; restore from transaction log; use DBCC CHECKDB; use third-party tools. Please back up your database regularly and enable transaction logging to prevent data loss.

SQL Server database files are usually stored in the following default location: Windows: C:\Program Files\Microsoft SQL Server\MSSQL\DATALinux: /var/opt/mssql/data The database file location can be customized by modifying the database file path setting.

When the SQL Server service fails to start, here are some steps to resolve: Check the error log to determine the root cause. Make sure the service account has permission to start the service. Check whether dependency services are running. Disable antivirus software. Repair SQL Server installation. If the repair does not work, reinstall SQL Server.

If the SQL Server installation fails, you can clean it up by following these steps: Uninstall SQL Server Delete registry keys Delete files and folders Restart the computer

The problem that SQL Server cannot be reinstalled due to incomplete deletion can be solved by following the following steps: manually delete files and registry entries; use SQL Server installation and uninstall tools; use third-party uninstall tools; check Windows Event Viewer; restart the computer; reinstall SQL Server.
