Linq 学习(6) Group & Join

本篇介绍Linq的Group和Join操作，继续使用《Linq 学习(3) 语法结构》中介绍的数据源。

Group
Group是进行分组操作，同SQL中的Group By类似。
原型如下：

public static IEnumerable<IGrouping<TKey, TSource>> GroupBy<TSource, TKey>(
    this IEnumerable<TSource> source,
    Func<TSource, TKey> keySelector)

它有几个重载，返回类型有两种：IEnumerable<IGrouping<TKey, TSource>> 和 IEnumerable<TResult>。
返回类型为：IEnumerable<IGrouping<TKey, TSource>>
示例：
返回按学号分组学生的成绩

var result = from score in DataSource.Scores
             group score by score.StudentID into scoreGroup
             select scoreGroup;

scoreGroup为IGrouping<TKey, TSource>类型，返回结果为IEnumerable<IGrouping<TKey, TSource>>，既集合的集合，因此输出时需用双重循环。
IGrouping<TKey, TElement>接口定义为：

public interface IGrouping<TKey, TElement> : IEnumerable<TElement>, IEnumerable
{
    TKey Key { get; }
}

其中Key为分组依据的字段。

foreach (var group in result)
{
    //输出分组依据的字段
    Console.WriteLine("\nStudent ID:" + group.Key);

    // 输出组内成员
    foreach (var score in group)
    {
        Console.WriteLine(score);
    }
}

// result:
// Student ID:1
// Student ID:1,Course ID:1,Score:78
// Student ID:1,Course ID:2,Score:60
// ...

// Student ID:2
// Student ID:2,Course ID:1,Score:59
// ...

等效的扩展方法调用实现为：

var result = DataSource.Scores.GroupBy(score => score.StudentID);

返回类型为：IEnumerable<TResult>
对分组结果进行一些包装，如包装为匿名类型。
返回按学号分组学生的成绩

var result = from score in DataSource.Scores
             group score by score.StudentID into scoreGroup
             select new { StudentID = scoreGroup.Key, Group = scoreGroup };

匿名类型中Group为IGrouping<TKey, TSource>类型。
等效的扩展方法调用实现为：

var result = DataSource.Scores.GroupBy(score => score.StudentID,
    (key, group) => new { StudentID = key, Group = group });

其他一些重载使用方法类似。

Join

连接操作。

public static IEnumerable<TResult> Join<TOuter, TInner, TKey, TResult>(
    this IEnumerable<TOuter> outer,
    IEnumerable<TInner> inner,
    Func<TOuter, TKey> outerKeySelector,
    Func<TInner, TKey> innerKeySelector,
    Func<TOuter, TInner, TResult> resultSelector)

从Join方法原型可以看出其使用方法。

内连接
选择左右两侧集合都含有相对应的元素。
示例：
查询学生的姓名、学科、成绩。

var result = from score in DataSource.Scores
             join student in DataSource.Students on score.StudentID equals student.StudentID
             join course in DataSource.Courses on score.CourseID equals course.CourseID
             select new { StudentName = student.Name, CourseName = course.CourseName, ScoreValue = score.Value };
// result
// { StudentName = Andy, CourseName = C Language, ScoreValue = 78 }
// { StudentName = Andy, CourseName = Biophysics, ScoreValue = 60 }
// ...
// { StudentName = Bill, CourseName = C Language, ScoreValue = 59 }
// { StudentName = Cindy, CourseName = Biophysics, ScoreValue = 60 }
// ...

等效的扩展方法调用实现为：

var result =
    DataSource.Scores.Join(
    DataSource.Students,
    score => score.StudentID,
    student => student.StudentID,
    (score, student) => new { StudentName = student.StudentID, ScoreValue = score.Value, CourseID = score.CourseID })
    .Join(DataSource.Courses,
    scostu => scostu.CourseID,
    course => course.CourseID,
    (scostu, course) => new { StudentName = scostu.StudentName, CourseName = course.CourseName, ScoreValue = scostu.ScoreValue }); 

左外连接
当右侧的连接的右侧没有左侧对应的元素时，内连接会忽略左侧元素。要想保留左侧元素，可以使用做外连接。右侧被置为默认值，如：引用类型被置为空。
示例：

var result =
    from student in DataSource.Students2
    join score in DataSource.Scores on student.StudentID equals score.StudentID into Scores
    from score in Scores.DefaultIfEmpty()
    select new { student = student, score = score == default(Score) ? 0 : score.Value };
// result:
// { student = Student ID:5,Student Name:Erik, score = 78 }
// { student = Student ID:6,Student Name:Frank, score = 0 }

等效的扩展方法调用实现为：
var result =
    DataSource.Students2.GroupJoin(
    DataSource.Scores,
    student => student.StudentID,
    score => score.StudentID,
    (student, Scores) => new { student = student, Scores = Scores })
    .SelectMany(group => group.Scores.DefaultIfEmpty(),
    (group, score) => new { student = group.student, score = (score == null) ? 0.0 : score.Value });

笛卡尔积
集合中的元素交错连接。
示例：统计学生课程成绩时的模板。

var result = from student in DataSource.Students
             from course in DataSource.Courses
             select new { StudentName = student.Name, CourseName = course.CourseName, ScoreValue = (double?)null };
// result:
// { StudentName = Andy, CourseName = C Language, ScoreValue =  }
// { StudentName = Andy, CourseName = Biophysics, ScoreValue =  }
// ...
// { StudentName = Bill, CourseName = C Language, ScoreValue =  }
// ...
// { StudentName = Cindy, CourseName = Fundamentals of Compiling, ScoreValue =  }
// ...

等效的扩展方法调用实现为：
var result = DataSource.Students.SelectMany(
    student => DataSource.Courses
        .Select(
        course =>
            new { StudentName = student.Name, CourseName = course.CourseName, ScoreValue = (double?)null }));

GroupJoin
连接分组。
方法原型为： 

public static IEnumerable<TResult> GroupJoin<TOuter, TInner, TKey, TResult>(
    this IEnumerable<TOuter> outer,
    IEnumerable<TInner> inner,
    Func<TOuter, TKey> outerKeySelector,
    Func<TInner, TKey> innerKeySelector,
    Func<TOuter, IEnumerable<TInner>, TResult> resultSelector)

// result:
// Andy
// 1----78
// 2----60
// ...
// Bill
// 1----59
// ...
// Cindy
// 2----60
// ...

相当于组合了Group操作和Join操作。等效的操作如下：

var result = from item in
                 (from student in DataSource.Students
                  join score in DataSource.Scores on student.StudentID equals score.StudentID
                  select new { StudentName = student.Name, CourseID = score.CourseID, Value = score.Value })
             group item by item.StudentName into Group
             select new { StudentName = Group.Key, Group = Group };

结束语
到现在，Linq与SQL语言等价的操作基本介绍完，组合这些操作能实现复杂的查询。