当我索引列时,为什么我的视图会变慢？

Question

在SQL Server中我已经把一个聚集索引上着眼于消除了低效率的连接使用LIKE语句:

   CREATE VIEW KeywordCount WITH SCHEMABINDING
    AS
    SELECT 
        K.ID AS KeywordID
        ,COUNT_BIG(*) AS KeywordCount
    FROM dbo.Grants G
    INNER JOIN dbo.GrantStatuses GS2 ON GS2.ID = G.StatusID AND GS2.Status NOT IN ('Pre-Submission', 'Awaiting Signatory Approval', 'Modifying', 'Closed')
    INNER JOIN dbo.Keywords K
        ON G.Keywords LIKE '%' + K.Word + '%'                                           --It's one of the grant's keywords
        OR G.Title LIKE '%' + K.Word + '%'                                              --Word appears in the title
        OR Replace(G.Title, '-', ' ') LIKE '%' + Replace(K.Word, '-', ' ') + '%'        --Word with hyphens replaced appears in the title
        OR G.Synopsis LIKE '%' + K.Word  + '%'                                          --Word appears in the Synopsis
        OR Replace(G.Synopsis, '-', ' ') LIKE '%' + Replace(K.Word, '-', ' ')+ '%'      --Word with hyphens replaced appears in the synopsis
    GROUP BY K.ID
    GO

    CREATE UNIQUE CLUSTERED INDEX IX_KeywordCount 
        ON dbo.KeywordCount (KeywordID)
    GO

然后我在KeywordCount列上添加了另一个索引:

    CREATE INDEX IX_KeywordCount_Count 
        ON dbo.KeywordCount (KeywordCount)
    GO

那么为什么以下查询需要7分钟才能运行？索引不应该给我更好的表现吗？

    SELECT TOP 10 * FROM KeywordCount ORDER BY KeywordCount DESC

编辑 谢谢大家,但我知道LIKE语句和REPLACE会使这个视图效率低下.这就是我添加聚集索引的原因.我认为将聚簇索引放在视图上会将数据实现到表中,这样数据库就不必进行连接.查询计划确实说它正在进行连接.这是为什么？

Answer 1

我在本文中找到了解决方案:http://technet.microsoft.com/en-us/library/cc917715.aspx

SELECT TOP 10 * FROM KeywordCount WITH (NOEXPAND) ORDER BY KeywordCount DESC

由于某种原因,查询计划没有使用索引,但我添加了WITH(NOEXPAND)提示,我的查询立即运行 - 非常感谢Quassnoi指出正确的事情要做.