分页SELECT UNION超级慢并杀死我的服务器

时间:2012-04-18 17:53:35

标签: sql-server sql-server-2008 query-optimization

我有一个SP,它返回包含UNION的查询中的分页数据。这会杀死我的数据库并且有时需要30秒才能运行,我在这里错过了一些明显的东西吗?我该怎么做才能提高它的表现?

参与的表:产品,类别,分类产品

目标: 任何不在类别中或已从类别UNION中删除的所有产品当前属于某个类别的所有产品,并通过它们覆盖网络服务。

我在所有列上都有索引,数据库中有427,996个产品,6148个类别和409,691个CategoryProducts。

这是我的查询,运行时间为6到30秒:

    SELECT * FROM (
    SELECT ROW_NUMBER() OVER(ORDER BY Products.ItemID, Products.ManufacturerID) AS RowNum, *
        FROM
        (
        SELECT  Products.*,
              CategoryID = NULL, CategoryName = NULL, 
              CategoryProductID = NULL, 
                          ContainerMinimumQuantity = 
                            CASE  COALESCE(Products.ContainerMinQty, 0)
                                WHEN 0 THEN Products.OrderMinimumQuantity
                                ELSE Products.ContainerMinQty
                            END 
                            Products.IsDeleted, 
                           SortOrder = NULL
    FROM    CategoryProducts RIGHT OUTER JOIN   Products 
        ON CategoryProducts.ManufacturerID = Products.ManufacturerID 
            AND CategoryProducts.ItemID = Products.ItemID
            WHERE     (Products.ManufacturerID = @ManufacturerID)
                AND (Products.ModifiedOn > @tStamp )
                 AND  ((CategoryProducts.IsDeleted = 1) OR (CategoryProducts.IsDeleted IS NULL))

                UNION

                SELECT Products.*, 
 CategoryProducts.CategoryID , CategoryProducts.CategoryName, 
          CategoryProducts.CategoryProductID , 
                          ContainerMinimumQuantity = 
                            CASE  COALESCE(Products.ContainerMinQty, 0)
                                WHEN 0 THEN Products.OrderMinimumQuantity
                                ELSE Products.ContainerMinQty
                            END 
                            CategoryProducts.IsDeleted,
                           CategoryProducts.SortOrder
    FROM      Categories INNER JOIN
                          CategoryProducts ON Categories.CategoryID = CategoryProducts.CategoryID INNER JOIN
                          Products ON CategoryProducts.ManufacturerID = Products.ManufacturerID
             AND CategoryProducts.ItemID = Products.ItemID
    WHERE     (Products.ManufacturerID = @ManufacturerID) 
            AND  (Products.ModifiedOn > @tStamp OR  CategoryProducts.ModifiedOn > @tStamp))  
            AS Products) AS C 
        WHERE RowNum >= @StartRow AND RowNum <= @EndRow

非常感谢任何见解。

1 个答案:

答案 0 :(得分:0)

如果我正确地阅读了您的情况,有两个不同查询的唯一原因是对缺失/删除​​的CategoryProducts的处理。我试图通过左边连接IsDeleted = 0解决这个问题,将所有已删除的CategoryProducts都置为空,所以我不必再次测试它们。 ModifiedOn部分为您想要检索的缺失/删除​​的Categoryproducts进行了另一个null测试。

select *
from (
    SELECT 
      Products.*, 
      -- Following three columns will be null for deleted/missing categories
      CategoryProducts.CategoryID, 
      CategoryProducts.CategoryName, 
      CategoryProducts.CategoryProductID , 
      ContainerMinimumQuantity = COALESCE(nullif(Products.ContainerMinQty, 0),
                                          Products.OrderMinimumQuantity),
      CategoryProducts.IsDeleted,
      CategoryProducts.SortOrder,
      ROW_NUMBER() OVER(ORDER BY Products.ItemID, 
                                 Products.ManufacturerID) AS RowNum
    FROM Products 
      LEFT JOIN CategoryProducts 
         ON CategoryProducts.ManufacturerID = Products.ManufacturerID
        AND CategoryProducts.ItemID = Products.ItemID
    -- Filter IsDeleted in join so we get nulls for deleted categories
    -- And treat them the same as missing ones
        AND CategoryProducts.IsDeleted = 0
      LEFT JOIN Categories
         ON Categories.CategoryID = CategoryProducts.CategoryID 
    WHERE Products.ManufacturerID = @ManufacturerID
      AND (Products.ModifiedOn > @tStamp 
        -- Deleted/missing categories
           OR CategoryProducts.ModifiedOn is null
           OR CategoryProducts.ModifiedOn > @tStamp)
    ) C
WHERE RowNum >= @StartRow AND RowNum <= @EndRow

在第三次看,除了作为CategoryProducts的过滤器之外,我没有看到任何类别被使用。如果是这种情况,则应将第二个LEFT JOIN更改为INNER JOIN,并将此部分括在括号中。