在LINQ中对IEnumerable进行排序[重复]

Question

可能重复：
Sorting an IEnumerable in LINQ

是否可以不使用 .ToList() 对其进行排序？

请参阅下面的代码。

下面的代码将产生这个输出。

    { id = 1, name = "sample 1", list = {'a','f','d'}},         
    { id = 5, name = "sample 1", list = {'a','f','c'}},
    { id = 2, name = "sample 1", list = {'g','b'}},
    { id = 4, name = "sample 1", list = {'i','e','h'}},
    { id = 6, name = "sample 1", list = {'d','b','c'}},
    { id = 3, name = "sample 1", list = {'h','i','c'}},

谢谢

RJ

IEnumerable<extra> eList = new List<extra>()
{
    new extra{ id = 1, text = "a"},
    new extra{ id = 2, text = "g"},
    new extra{ id = 3, text = "i"},
    new extra{ id = 4, text = "e"},
    new extra{ id = 5, text = "f"},
    new extra{ id = 6, text = "d"},
    new extra{ id = 7, text = "c"},
    new extra{ id = 8, text = "h"},
    new extra{ id = 9, text = "b"}
};

IEnumerable<sample> sam = new List<sample>()
{
    new sample{ id = 1, name = "sample 1", list = new List<int>{1,5,6}},
    new sample{ id = 2, name = "sample 2", list = new List<int>{2,9}},
    new sample{ id = 3, name = "sample 3", list = new List<int>{8,3,7}},
    new sample{ id = 4, name = "sample 4", list = new List<int>{3,4,8}},
    new sample{ id = 5, name = "sample 5", list = new List<int>{1,5,7}},
    new sample{ id = 6, name = "sample 6", list = new List<int>{6,9,7}}
};

var sorted = (from d1 in sam
              select new
              {
                  name = d1.name,
                  id = d1.id,
                  list =
                  (
                      from d2 in d1.list
                      join e in eList on d2 equals e.id
                      select e.text
                  ).OrderBy(item => item).ToList()
              }).OrderBy(item => item.list.FirstOrDefault());

Answer 1

也许 ThenBy 会这样做？

var sorted = (from d1 in sam 
              select new 
              { 
                  name = d1.name, 
                  id = d1.id, 
                  list = 
                  ( 
                      from d2 in d1.list 
                      join e in eList on d2 equals e.id 
                      select e.text 
                  ).OrderBy(item => item 
              }).ThenBy(item => item.list.FirstOrDefault()); 

我误解了这个问题。为什么不直接删除 ToList？您可能希望使用 .ToList() 来防止每次访问集合时进行排序。

Answer 2

我刚刚测试了没有“ToList”的原始代码，它对项目和“附加”进行了很好的排序。你能详细说明你想要实现的目标吗？证明：

使用 ToList 的原始代码结果：

1  "sample 1"  a d f  
5  "sample 5"  a c f  
2  "sample 2"  b g  
6  "sample 6"  b c d  
3  "sample 3"  c h i  
4  "sample 4"  e h i  

没有 ToList 的代码结果：

1  "sample 1"  a d f   <-- adf is sorted
5  "sample 5"  a c f   <-- also sorted
2  "sample 2"  b g     <-- and here too
6  "sample 6"  b c d   <-- yep
3  "sample 3"  c h i   <-- it is!
4  "sample 4"  e h i   <-- ...
               ^
               |
               \ aabbce is sorted, too

跟我一模一样:)

至于一般的想法，在没有“ToList”的情况下，用 LINQ 语法编写的相同的小问题是 JOIN 将被懒惰地执行，每个项目一次，并且每次它都会从头开始构建映射/连接，而它是完全可缓存和可重用的。下面，在扩展语法中，我只显式地预先创建了一次映射，然后所有其他没有任何具体化的惰性查询使用共享相同的映射。这样它可能会快几倍并且使用更少的内存：

var mapping = eList.ToDictionary(x => x.id, x=>x.text);

var temps = sam.Select(s =>
    new {
        id = s.id,
        name = s.name,
        stringlist = s.list.Select(id => mapping[id]).OrderBy(str => str)
    });

var result = temps.OrderBy(t => t.stringlist.FirstOrDefault());

Answer 3

试试这个：

var sorted = (from d1 in sam
              select new
              {
                  name = d1.name,
                  id = d1.id,
                  list =
                  (
                      from d2 in d1.list
                      join e in eList on d2 equals e.id
                      select e.text
                  )
              }).OrderBy(item => item.list.OrderBy(i => i).FirstOrDefault());