SQL where 条件中的 IN 子句导致大数据出现性能问题

Question

以下查询在我的数据库中运行良好，但它在客户数据库中产生了巨大的性能问题。我知道我在 where 条件中使用了 IN 子句，这给我带来了这个问题。但我不知道如何解决这个问题。

declare @AccountId int
set @AccountId =  1200
declare @IsLinkedAccountsNotes bit
set @IsLinkedAccountsNotes =1
declare @EventType varchar(100)
set @EventType = ''

SELECT
        u.loginName as InteractionLoginName,
        u.userName as InteractionUserName,
        e.*
    FROM
        lat.events e
        INNER JOIN dbo.MasterEvents me ON me.EventId = e.EventId    
        LEFT JOIN dbo.Users u ON e.UserId = u.ID 
    WHERE
        (me.AccountId = @AccountId OR
        (@IsLinkedAccountsNotes = 1 AND me.AccountId IN (SELECT DISTINCT [number] FROM dbo.Linking_LinkedAccounts WHERE linked_number = @AccountId) AND e.EventType = 'AccountNoteAdded'))

我知道是 where 子句中的第二个条件导致了问题。而且我在各种帖子中看到使用连接可以解决这个问题。但是我不知道如何在 where 条件下使用 join 。或者有没有其他方法可以做到这一点。

请帮忙。

Answer 1

可以试试

JOIN (SELECT DISTINCT [number] FROM dbo.Linking_LinkedAccounts WHERE linked_number = @AccountId) lla ON lla.number = me.AccountId

LEFT JOIN 并过滤掉 where 条件中的空值。

EXISTS 而不是 IN。

Answer 2

您的问题很可能是 "catch all query"。您可以尝试按以下方式拆分案例

SELECT u.loginName AS InteractionLoginName,
       u.userName  AS InteractionUserName,
       e.*
FROM   lat.events e
       INNER JOIN dbo.MasterEvents me
               ON me.EventId = e.EventId
       LEFT JOIN dbo.Users u
              ON e.UserId = u.ID
WHERE  me.AccountId = @AccountId
UNION ALL
SELECT u.loginName AS InteractionLoginName,
       u.userName  AS InteractionUserName,
       e.*
FROM   lat.events e
       INNER JOIN dbo.MasterEvents me
               ON me.EventId = e.EventId
       LEFT JOIN dbo.Users u
              ON e.UserId = u.ID
WHERE  @IsLinkedAccountsNotes = 1
       AND e.EventType = 'AccountNoteAdded'
       AND me.AccountId IN (SELECT [number]
                            FROM   dbo.Linking_LinkedAccounts
                            WHERE  linked_number = @AccountId
                                   AND [number] <> @AccountId)

SQL where 条件中的 IN 子句导致大数据出现性能问题

SQL IN clause in where condition causing performance issue with large data

sql

sql-server-2008

query-performance

sql-server-2012