0

我有这个查询，它给了我需要但执行速度非常慢的结果。缓慢是因为 [Contoso$Item Ledger Entry] 表有 ~0.6M 记录。请建议如何优化/简化此查询（如果可能）以及我需要哪些索引才能使其发挥最佳效果。

DECLARE @TempTable TABLE 
(
    Numeris varchar(20)
)
INSERT INTO @TempTable (Numeris)
SELECT No_
FROM [Contoso$Sales Line]
WHERE [Document no_] = 'PP03428'

SELECT I.No_, CASE WHEN 
           ISNULL(ROUND(CAST(SUM(ILE.[Remaining Quantity]) AS float), 2), 0) <
           (SELECT  ISNULL(SUM([Outstanding Quantity]), 0)
            FROM    [Contoso$Sales Line] SL
            WHERE   (SL.No_ = I.No_) 
                AND (SL.[Document type] = 1 OR SL.[Document type] = 2) 
                AND (SL.[DA4 eksportas] = 0) 
                AND (SL.Type = 2)
           ) 
        THEN 0 
        ELSE 
           ISNULL(ROUND(CAST(SUM(ILE.[Remaining Quantity]) AS float), 2) -
           (SELECT  ISNULL(SUM([Outstanding Quantity]), 0)
            FROM    [Contoso$Sales Line] SL
            WHERE   (SL.No_ = I.No_) 
                AND (SL.[Document type] = 1 OR SL.[Document type] = 2) 
                AND (SL.[DA4 eksportas] = 0) 
                AND (SL.Type = 2)), 0
           ) 
    END AS Kiekis
FROM [Contoso$Item] I 
LEFT OUTER JOIN [Contoso$Item Ledger Entry] ILE ON I.No_ = ILE.[Item No_]
WHERE I.No_ IN (SELECT Numeris FROM @TempTable)
GROUP BY I.No_
4

1 回答 1

0

使用执行计划来了解您的查询是如何执行的，并知道您的查询为何如此缓慢。您可以在 Contoso$Item 表的 No_ 列上添加索引以避免全扫描表

于 2013-04-30T14:58:17.760 回答