Question

我有以下代码：

var comp = db.Companies.Select(x => new
        {

            topid = x.ID_comp,
            hill = ((from company in db.Companies
                    where x.ID_comp > company.ID_comp
                    orderby company.ID_comp
                    select new { h = company.ID_comp.ToString() })
                    .Concat((
                    from company in db.Companies
                    where x.ID_comp > company.ID_comp
                    orderby company.ID_comp descending
                    select new { h = company.ID_comp.ToString() })))
                    .Select(y => y.h)
        });

任务是为每个公司的ID生成ID的“金字塔”：121为3，12321为4，依此类推（如果存在id像1，2，3 4）。给定查询的想法是在2个集合中获取所有小于当前ID的所有ID，然后将它们连接起来，以降序对第二个ID排序（金字塔的顶部重复，但是我稍后会返回它）。 / p>

问题在于order by子句没有分别针对每个集合执行，最后对于ID为4的公司，我得到了类似1 2 3 1 2 3的信息。不同的订单？

UPD1：我使用的数据：

表db.Companies

ID_comp | name
--------------------
   1    |Don_avia  
   2    |Aeroflot  
   3    |Dale_avia 
   4    |air_France
   5    |British_AW

输出：

1:
2: 1 1
3: 1 2 1 2
4: 1 2 3 1 2 3
5: 1 2 3 4 1 2 3 4

UPD2：结果查询：

SELECT 
[Extent1].[ID_comp] AS [ID_comp], 
[Project3].[C2] AS [C1], 
[Project3].[C1] AS [C2]
FROM  [dbo].[Company] AS [Extent1]
OUTER APPLY  (SELECT 
    [UnionAll1].[C1] AS [C1], 
    1 AS [C2]
    FROM  (SELECT 
         CAST( [Extent2].[ID_comp] AS nvarchar(max)) AS [C1]
        FROM [dbo].[Company] AS [Extent2]
        WHERE [Extent1].[ID_comp] > [Extent2].[ID_comp]
    UNION ALL
        SELECT 
         CAST( [Extent3].[ID_comp] AS nvarchar(max)) AS [C1]
        FROM [dbo].[Company] AS [Extent3]
        WHERE [Extent1].[ID_comp] > [Extent3].[ID_comp]) AS [UnionAll1] ) AS [Project3]
ORDER BY [Extent1].[ID_comp] ASC, [Project3].[C2] ASC

Answer 1

好像您遇到了EF6查询翻译错误-.Concat（SQL UNION ALL）子查询的顺序将被忽略。

我发现的解决方法是将伪造的限制运算符（如Take(int.MaxValue)附加到子查询中，这导致EF6考虑排序：

hill =
    (from company in db.Companies
     where x.ID_comp > company.ID_comp
     orderby company.ID_comp
     select new { h = company.ID_comp.ToString() }
    ).Take(int.MaxValue)
    .Concat(
    (from company in db.Companies
     where x.ID_comp > company.ID_comp
     orderby company.ID_comp descending
     select new { h = company.ID_comp.ToString() }
    ).Take(int.MaxValue))
    .Select(y => y.h)

生成的SQL：

SELECT
    [Extent1].[ID_comp] AS [ID_comp],
    [Project5].[C2] AS [C1],
    [Project5].[C1] AS [C2]
    FROM  [dbo].[Company] AS [Extent1]
    OUTER APPLY  (SELECT
        [UnionAll1].[C1] AS [C1],
        1 AS [C2]
        FROM  (SELECT TOP (2147483647)
            [Project1].[C1] AS [C1]
            FROM ( SELECT
                [Extent2].[ID_comp] AS [ID_comp],
                 CAST( [Extent2].[ID_comp] AS nvarchar(max)) AS [C1]
                FROM [dbo].[Company] AS [Extent2]
                WHERE [Extent1].[ID_comp] > [Extent2].[ID_comp]
            )  AS [Project1]
            ORDER BY [Project1].[ID_comp] ASC
        UNION ALL
            SELECT TOP (2147483647)
            [Project3].[C1] AS [C1]
            FROM ( SELECT
                [Extent3].[ID_comp] AS [ID_comp],
                 CAST( [Extent3].[ID_comp] AS nvarchar(max)) AS [C1]
                FROM [dbo].[Company] AS [Extent3]
                WHERE [Extent1].[ID_comp] > [Extent3].[ID_comp]
            )  AS [Project3]
            ORDER BY [Project3].[ID_comp] DESC) AS [UnionAll1] ) AS [Project5]
    ORDER BY [Extent1].[ID_comp] ASC, [Project5].[C2] ASC

Answer 2

var comp = db.Companies.Select(x => new
{
    topid = x.ID_comp,
    hill = ((from company in db.Companies.OrderBy(y => y.ID_comp)
        where x.ID_comp > company.ID_comp
        select new { h = company.ID_comp.ToString() })
        .Concat((
        from company in db.Companies.OrderByDescending(y => y.ID_comp)
        where x.ID_comp > company.ID_comp
        select new { h = company.ID_comp.ToString() })))
        .Select(y => y.h)
    });

LINQ Concat嵌套查询排序

2 个答案: