使用LINQ to实体生成信息

在第一部分,有趣的限制是书必须匹配输入的每个标记,所以“where tagid==someid”的where子句不会真正起作用。我设想类似这样的事情(Linq to Objects示例)

List<int> selectedTagIds = new List<int>() { 1, 2 };
var query = from book in books
            join booktag in booktags 
            on book.Id equals booktag.BookId 
            join selectedId in selectedTagIds 
            on booktag.TagId equals selectedId 
            group book by book into bookgroup 
            where bookgroup.Count() == selectedTagIds.Count
            select bookgroup.Key;

它基本上执行从书籍到书签以及所选标记ID列表的联接,并将所选内容限制为书籍计数与所选标记ID计数匹配的位置。

去拉相关的标签,也许是这样的

var relatedTags = from book in query // use original query as base
                    join booktag in booktags
                    on book.Id equals booktag.BookId
                    join tag in tags
                    on booktag.TagId equals tag.Id
                    where !selectedTagIds.Contains(tag.Id) // exclude selected tags from related tags
                    group tag by tag into taggroup
                    select new
                    {
                        Tag = taggroup.Key,
                        Count = taggroup.Count()
                    };

快速示例的完整代码。不是很糟糕,但你明白了。

using System;
using System.Collections.Generic;
using System.Linq;

namespace StackOverflow
{
    class Program
    {
        static void Main()
        {
            List<Book> books = new List<Book>() 
            {
                new Book() { Id = 113421, Title = "A" },
                new Book() { Id = 113422, Title = "B" }
            };

            List<Tag> tags = new List<Tag>()
            {
                new Tag() { Id = 1, Name = "ASP" },
                new Tag() { Id = 2, Name = "C#" },
                new Tag() { Id = 3, Name = "CSS" },
                new Tag() { Id = 4, Name = "VB" },
                new Tag() { Id = 5, Name = "VB.NET" },
                new Tag() { Id = 6, Name = "PHP" },
                new Tag() { Id = 7, Name = "Java" },
                new Tag() { Id = 8, Name = "Pascal" }
            };

            List<BookTag> booktags = new List<BookTag>()
            {
                new BookTag() { Id = 1, BookId = 113421, TagId = 1 },
                new BookTag() { Id = 2, BookId = 113421, TagId = 2 },
                new BookTag() { Id = 3, BookId = 113421, TagId = 3 },
                new BookTag() { Id = 4, BookId = 113421, TagId = 4 },
                new BookTag() { Id = 5, BookId = 113422, TagId = 1 },
                new BookTag() { Id = 6, BookId = 113422, TagId = 4 },
                new BookTag() { Id = 7, BookId = 113422, TagId = 8 }
            };


            List<int> selectedTagIds = new List<int>() { 1,2 };

            // get applicable books based on selected tags

            var query = from book in books
                        join booktag in booktags
                        on book.Id equals booktag.BookId
                        join selectedId in selectedTagIds
                        on booktag.TagId equals selectedId
                        group book by book into bookgroup
                        where bookgroup.Count() == selectedTagIds.Count
                        select bookgroup.Key;

            foreach (Book book in query)
            {
                Console.WriteLine("{0}\t{1}",
                    book.Id,
                    book.Title);
            }

            // get related tags for selected tags

            var relatedTags = from book in query // use original query as base
                              join booktag in booktags
                              on book.Id equals booktag.BookId
                              join tag in tags
                              on booktag.TagId equals tag.Id
                              where !selectedTagIds.Contains(tag.Id) // exclude selected tags from related tags
                              group tag by tag into taggroup
                              select new
                              {
                                  Tag = taggroup.Key,
                                  Count = taggroup.Count()
                              };

            foreach (var relatedTag in relatedTags)
            {
                Console.WriteLine("{0}\t{1}\t{2}",
                    relatedTag.Tag.Id,
                    relatedTag.Tag.Name,
                    relatedTag.Count);
            }

            Console.Read();
        }
    }

    class Book
    {
        public int Id { get; set; }
        public string Title { get; set; }
    }

    class Tag
    {
        public int Id { get; set; }
        public string Name { get; set; }
    }

    class BookTag
    {
        public int Id { get; set; }
        public int BookId { get; set; }
        public int TagId { get; set; }
    } 
}

因此,对于选定的标记1&2,您将得到BookA,相关的标记将是3(css)和4(vb)。

只需使用外键映射1:n或1:1关系中的表,并让设计器为您创建导航属性。(books:books tags映射1:n从books.bookid到books tags.bookid,books tags.tagid映射1:1到tags.tagid)。这实际上是一种伪装的N:M关系。我不知道设计师是否直接接受了这个,但是通过一些修改,你可以得到正确的导航属性。

现在请回答以下问题:

1. model.Tags.Where(t => t.ID == 1).Books.Select(b => b.ID)

2. 获取一本书的所有标签,并将该表加入到bookstags中,通过这个,您可以简单地使用count()来获取计数。

这与答案没有直接关系,但您可能想看看 linqpad 因为它将帮助您直接从数据库构建L2S语句。