Question

我有以下查询在Sql Server 2008中运行，如果数据很小，它可以很好地工作但是当它很大时我得到了异常。有什么方法可以优化查询

select
        distinct olu.ID
            from olu_1 (nolock) olu

            join mystat (nolock) s
                on s.stat_int = olu.stat_int

            cross apply
                dbo.GetFeeds 
                        (
                                s.stat_id,
                                olu.cha_int, 
                                olu.odr_int,  
                                olu.odr_line_id, 
                                olu.ID
                        ) channels

            join event_details (nolock) fed
                on fed.air_date = olu.intended_air_date
                and fed.cha_int = channels.cha_int
                and fed.break_code_int = olu.break_code_int

            join formats (nolock) fmt
                on fed.format_int = fmt.format_int


            where
                fed.cha_int in (125, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 35, 36, 37, 38, 39, 40, 41, 43, 117, 45, 42, 44, 47, 49, 50, 51, 46, 52, 53, 54, 55, 56, 48, 59, 60, 57, 63, 58, 62, 64, 66, 69, 68, 67, 65, 70, 73, 71, 74, 72, 75, 76, 77, 78, 79, 82, 80, 159, 160, 161, 81, 83, 84, 85, 88, 87, 86, 89, 90, 61, 91, 92, 93, 95, 96, 97, 98, 99, 100, 94, 155, 156, 157, 158, 103, 104, 102, 101, 105, 106, 107, 108, 109, 110, 119, 111, 167, 168, 169, 112, 113, 114, 115, 116, 170, 118, 120, 121, 122, 123, 127, 162, 163, 164, 165, 166, 128, 129, 130, 124, 133, 131, 132, 126, 134, 136, 135, 137, 171, 138, 172, 173, 174) and
                fed.air_date between '5/27/2013 12:00:00 AM' and '6/2/2013 12:00:00 AM' and

                fmt.cha_int in (125, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151, 152, 153, 154, 35, 36, 37, 38, 39, 40, 41, 43, 117, 45, 42, 44, 47, 49, 50, 51, 46, 52, 53, 54, 55, 56, 48, 59, 60, 57, 63, 58, 62, 64, 66, 69, 68, 67, 65, 70, 73, 71, 74, 72, 75, 76, 77, 78, 79, 82, 80, 159, 160, 161, 81, 83, 84, 85, 88, 87, 86, 89, 90, 61, 91, 92, 93, 95, 96, 97, 98, 99, 100, 94, 155, 156, 157, 158, 103, 104, 102, 101, 105, 106, 107, 108, 109, 110, 119, 111, 167, 168, 169, 112, 113, 114, 115, 116, 170, 118, 120, 121, 122, 123, 127, 162, 163, 164, 165, 166, 128, 129, 130, 124, 133, 131, 132, 126, 134, 136, 135, 137, 171, 138, 172, 173, 174) and
                fmt.air_date between '5/27/2013 12:00:00 AM' and '6/2/2013 12:00:00 AM'

Answer 1

来自IN (Transact-SQL)

包含极大数量的值（数千） IN子句可以消耗资源并返回错误8623或8632. To 解决此问题，将项目存储在表格的IN列表中。

所以我建议将值插入到临时表中，然后连接到该表或从表中选择IN

类似

DECLARE @TABLE TABLE(
        val INT
)

INSERT INTO @TABLE VALUES(1),(2),(3),(4),(5)

SELECT  *
FROM    MyTable
WHERE   ID IN (SELECT val FROM @TABLE)

SQL Fiddle DEMO

Answer 2

从您的生产数据库（包含许多行）进行备份，并在您的开发计算机上本地使用它。优化可能需要一些时间，如果你是sql的新手可能实际上很难。将Query分解为几个临时表，并最终将它们连接起来。尝试从查询中删除dbo.GetFeeds（...）函数以查看该函数是否存在问题。

查询处理器用完了内部资源异常

2 个答案:

SQL Fiddle DEMO