单个查询在Entity Framework中生成> 1100个不同的查询

Question

我有一个奇怪的问题 - 我有一个名为Item的表,它通过Item表中的BrandID键链接到一个名为brand的表.

我有一个EF Code First模型,它在Items表中具有如下的Navigation属性:

public virtual Brand Brand { get; set; }

我有一个项目列表,我发送给一个名为FilterItems的方法,该方法从另一个方法获取谓词,执行Where语句以返回已过滤图标的列表:

   public IEnumerable<Item> FilterItems(IEnumerable<Item> items)
            {
                string searchString = txtSearchString.Text.ToUpper();

                Func<Item, bool> predicate;

                switch (Field)
                {
                    case SearchField.Brand:
                    default:
                        predicate = BuildBrandPredicate(searchString);
                        break;
                    case SearchField.Model:
                        predicate = BuildModelPredicate(searchString);
                        break;
                    case SearchField.Note:
                        predicate = BuildNotePredicate(searchString);
                        break;
                    case SearchField.Reference:
                        predicate = BuildReferencePredicate(searchString);
                        break;
                    case SearchField.Text:
                        predicate = BuildTextPredicate(searchString);
                        break;
                }

                var result = items.Where(predicate);
                return result;
            }

        private Func<Item, bool> BuildBrandPredicate(string searchString)
        {
            Func<Item, bool> predicate;
            //build the predicate for brand
            switch (cboSearchActions.Text)
            {
                case "Exact":
                    predicate = (item => item.Brand.Description.ToUpper() == searchString);
                    break;
                //Other similar functions go here but I am concentrating on Exact
            }
            return predicate;
        }

数据库中有大约32000个项目和1000个品牌,每个项目仅链接到一个品牌.

搜索非常慢,当我调试SQL时,我发现它为品牌表中的每个记录运行此序列:

Opened connection at 29/09/2014 15:14:46 +01:00

SELECT 
    [Extent1].[ID] AS [ID], 
    [Extent1].[Description] AS [Description]
    FROM [Brand] AS [Extent1]
    WHERE [Extent1].[ID] = @EntityKeyValue1


-- EntityKeyValue1: '1' (Type = Int32, IsNullable = false)

-- Executing at 29/09/2014 15:14:46 +01:00

-- Completed in 6 ms with result: SqlCeDataReader

这共运行了1123次,这很荒谬.

当然生成的sql应该是一个带有内连接的单个sql语句？

任何人都可以解释为什么会发生这种情况以及我是否可以做些什么来阻止这种荒谬的行为

我在用

1. Visual Studio 2012
2. C#
3. Sql Server Compact 4.0
4. 实体框架6
5. .Net 4.5

Answer 1

IEnumerable<T>是对象的linq - 你告诉它每个项目单独执行这些操作.除非您使用LINQ-to-some-backend,例如LINQ-to-Entities,否则它不可能组成查询.幸运的是,这通常是在更换简单IEnumerable<T>用IQueryable<T>,并Func<Foo,Bar>具有Expression<Func<Foo,Bar>>:

    public IQueryable<Item> FilterItems(IQueryable<Item> items)
    {
        string searchString = txtSearchString.Text.ToUpper();

        Expression<Func<Item, bool>> predicate;

        switch (Field)
        {
            case SearchField.Brand:
            default:
                predicate = BuildBrandPredicate(searchString);
                break;
            case SearchField.Model:
                predicate = BuildModelPredicate(searchString);
                break;
            case SearchField.Note:
                predicate = BuildNotePredicate(searchString);
                break;
            case SearchField.Reference:
                predicate = BuildReferencePredicate(searchString);
                break;
            case SearchField.Text:
                predicate = BuildTextPredicate(searchString);
                break;
        }

        var result = items.Where(predicate);
        return result;
    }
    private Expression<Func<Item, bool>> BuildBrandPredicate(string searchString)
    {
        Expression<Func<Item, bool>> predicate;
        //build the predicate for brand
        switch (cboSearchActions.Text)
        {
            case "Exact":
                predicate = (item => item.Brand.Description.ToUpper() == searchString);
                break;
            //Other similar functions go here but I am concentrating on Exact
        }
        return predicate;
    }