Question

我使用此语句为excel创建列表

SELECT DISTINCT Year, Version 
    FROM myView 
    WHERE id <> 'old' 
    ORDER BY Year DESC, Version DESC

问题是执行时间超过30秒，因为有近200万行。

结果只有大约1000行。

为了加快执行时间，我有哪些选项只提取这两列？我还需要确保识别基础表的插入。

我是否需要新表来复制视图中的值？还有一个管理更新的触发器？

谢谢

Answer 1

所以，大概有一张桌子上有年份和身份证明你的观点。鉴于这个（微不足道的）例子：

CREATE TABLE myTable ([id] varchar(10), [Year] int, [Version] int);

只需在该表上创建一个与您查询数据的方式相匹配的索引。鉴于您的查询：

SELECT DISTINCT Year, Version 
  FROM myView 
 WHERE id <> 'old' 
 ORDER BY Year DESC, Version DESC

此查询与WHERE和ORDER BY子句匹配，并且应该为您提供所需的所有性能：

IF  EXISTS (SELECT * FROM sys.indexes WHERE object_id = OBJECT_ID(N'[dbo].[myTable]') AND name = N'IX_YearVersion_Filtered')
    DROP INDEX [IX_YearVersion_Filtered] ON [dbo].[myTable] WITH ( ONLINE = OFF )
GO
CREATE NONCLUSTERED INDEX [IX_YearVersion_Filtered] ON [dbo].[myTable] 
(
    [Year] DESC,
    [Version] DESC
)
WHERE ([id]<>'old')
GO

Answer 2

with cte_x
as
(SELECT Year, Version 
    FROM myView 
    WHERE id not in ('old')
    group by Year, Version)
SELECT DISTINCT Year, Version 
    FROM cte_x
        ORDER BY Year DESC, Version DESC

TSQL - 将数据提取到表/视图以加速查询

2 个答案: