我首先使用实体框架6代码,我有一个简单的模型:
public class Task
{
[Key]
public int aid {get;set;}
[MaxLength(256)]
public string Memo {get;set;}
}
我得到了一个模特:
int id = 3;
from t in db.Tasks
where t.aid == id
select t;
或
int id = 3;
db.Tasks.Find(id);
肯定会很快但不是......
我查看由EF ORM生成的IntelliTrace中的SQL,如下所示:
DECLARE @p__linq__0 AS SQL_VARIANT;
SET @p__linq__0 = 3;
SET STATISTICS TIME ON
SET STATISTICS IO ON
SELECT
[Limit1].[aid] AS [aid],
[Limit1].[Memo] AS [Memo]
FROM ( SELECT TOP (1)
[Extent1].[aid] AS [aid],
[Extent1].[Memo] AS [Memo]
FROM [dbo].[Task] AS [Extent1]
WHERE [Extent1].[aid] = @p__linq__0
) AS [Limit1]
SET STATISTICS TIME OFF
SET STATISTICS IO OFF
我添加SET STATISTICS并在SSMS中测试它。
表'任务'。 扫描计数1 ,...
使用 SQL_VARIANT !执行计划是扫描表而不是集群搜索!
为什么EF会这样做?!我可以避免吗?
(使用SQL Server 2012的LocalDB)
答案 0 :(得分:4)
我发现了一个问题!
它的原因我在IntelliTrace中查看sql 。
IntelliTrace将隐藏所有变体并将其显示为为SQL_VARIANT 。
我通过SQL Server Profiler获得了实际的SQL,sql是:
exec sp_executesql N'SELECT
[Limit1].[aid] AS [aid],
[Limit1].[Memo] AS [Memo]
FROM ( SELECT TOP (1)
[Extent1].[aid] AS [aid],
[Extent1].[Memo] AS [Memo]
FROM [dbo].[Task] AS [Extent1]
WHERE [Extent1].[aid] = @p__linq__0
) AS [Limit1]',N'@p__linq__0 int',@p__linq__0=3
可以使用寻求查询数据。