Question

我有一个SQL查询，大约需要1分钟才能获得75,000条记录。

我想知道是否有办法通过更改任何内容来加快速度。

        SELECT      c.CampaignID,
                    c.Description,
                    c.Active,
                    c.CampaignTypeID,
                    CASE c.CampaignTypeID 
                        WHEN 1 THEN 'TEST1'
                        WHEN 2 THEN 'TEST2'
                        WHEN 3 THEN 'TEST3'
                        WHEN 5 THEN 'TEST4'
                        WHEN 6 THEN 'TEST5'
                    END AS CampaignType,
                    CASE c.Active
                        WHEN 1 THEN 'TEST6'
                        ELSE 'TEST7'
                        END  AS style,
                    CASE c.Active
                        WHEN 1 THEN 'Deactivate' 
                        ELSE 'Activate'
                        END AS stat,
                    REPLACE( REPLACE(c.Description,'\\','\\\\'),'''','\\') AS LinkDescription,
                    @CompanyID AS CompanyID 
        FROM        Campaign c WITH(nolock)
                    LEFT OUTER JOIN UserPermissions 
                        ON up.CampaignID=c.CampaignID 
                        AND up.UserID = @UserID
        WHERE       c.CompanyID = @CompanyID
                    and c.IsCampaignHidden = 0
                    AND (@IsAdmin = 0
                    and case when (@IsAdmin = 0) then up.UserID else 0 end is not null
                    and case when (@IsAdmin = 0) then up.[View] else 1 end = 1

感谢任何帮助。

Answer 1

这令人困惑......

您的条件AND (@IsAdmin = 0后面跟着另一个条件case when (@IsAdmin = 0)，您已经定义了@IsAdmin = 0，我认为以下case when不是必需的，因为它始终转向{ {1}}因为您已经定义了TRUE。您必须清理代码，删除不需要的行，因为这会导致额外的执行时间，特别是在@IsAdmin = 0条件下。

您应避免在WHERE中使用多个CASE WHEN条件，因为这会导致执行时间过长。

WHERE

我认为你不需要WHERE c.CompanyID = @CompanyID and c.IsCampaignHidden = 0 AND (@IsAdmin = 0 and case when (@IsAdmin = 0) then up.UserID else 0 end is not null and case when (@IsAdmin = 0) then up.[View] else 1 end = 1

Answer 2

尝试删除

REPLACE( REPLACE(c.Description,'\\','\\\\'),'''','\\') AS LinkDescription

看看它运行了多长时间;如果它更快，找到一种不同的方式去做正在做的事情。

SQL查询1分钟即可获得75,000条记录

2 个答案: