有朋友反馈说我提供的sample不能编译。大概是版本的问题，可以到下载for beta1的版本。本节接着讲groupby。
上一节，我们讲了如何理解groupby返回的结果。本节会延这个思路阐述下去。先来看下面的例子 GroupBy操作中Select的匿名类
var q = from p in db.Products group p by p.CategoryID into g select new { CategoryID = g. Key , g }; 
如果，使用下面的语句。 var q = from p in db.Products group p by p.CategoryID into g select new { CategoryID = g. Key ,GroupSet = g }; foreach (var gp in q)
{ 
if (gp.CategoryID == 7)
{ foreach (var p in gp.GroupSet) { 
} }
} SELECT MAX ( [ t0 ] . [ UnitPrice ] ) AS [ MaxPrice ] , [ t0 ] . [ CategoryID ] FROM [ dbo ] . [ Products ] AS [ t0 ] GROUP BY [ t0 ] . [ CategoryID ] var q = from p in db.Products group p by p.CategoryID into g select new { g. Key , MaxPrice = g. Max (p => p.UnitPrice) }; 在这里，Max函数只对每个分组进行操作。我们来看看其结果
呀，这次，dlinq并没有把组里所有的纪录都取出来的吗。（请参考一文中的方法，配置sample.） dlinq只是简单做了统计，并返回结果。 每类产品中，单价为最小的， var q = from p in db.Products group p by p.CategoryID into g select new { g. Key , MinPrice = g. Min (p => p.UnitPrice) }; var q = from p in db.Products group p by p.CategoryID into g select new { g. Key , AveragePrice = g.Average(p => p.UnitPrice) }; var q = from p in db.Products group p by p.CategoryID into g select new { g. Key , TotalPrice = g. Sum (p => p.UnitPrice) }; var q = from p in db.Products group p by p.CategoryID into g select new { g. Key , NumProducts = g. Count () }; var q = from p in db.Products group p by p.CategoryID into g select new { g. Key , NumProducts = g. Count (p => p.Discontinued) }; var q = from p in db.Products group p by p.CategoryID into g where g. Count () >= 10 select new { g. Key , ProductCount = g. Count () }; SELECT [ t1 ] . [ CategoryID ] , [ t1 ] . [ value2 ] AS [ ProductCount ] FROM ( SELECT COUNT ( * ) AS [ value ] , COUNT ( * ) AS [ value2 ] , [ t0 ] . [ CategoryID ] FROM [ dbo ] . [ Products ] AS [ t0 ] GROUP BY [ t0 ] . [ CategoryID ] ) AS [ t1 ] WHERE [ t1 ] . [ value ] >= @p0 -- @p0: Input Int32 (Size = 0; Prec = 0; Scale = 0) [10] -- Context: SqlProvider(Sql2005) Model: AttributedMetaModel Build: 2.0.20612.0 GroupBy操作中GroupBy的匿名类在时，我们就提到不光Select操作可以使用匿名类，其他操作符也可以。但是，OrderBy不支持。请参考
当用户既想按产品的分类，又想按供应商来做分组，该怎么办呢。这时，我们就该使用匿名类。 var categories = from p in db.Products group p by new { p.CategoryID, p.SupplierID } into g select new {g. Key , g}; 在by后面，new出来一个匿名类。这里，Key其实质是一个类的对象，Key包含两个Property,一个是CategoryID，再一个是SupplierID ，要想取到具体CategoryID的值，需要g.Key.CategoryID，才能访问到。我们来看dlinq翻译的T-sql语句。
SELECT [ t0 ] . [ SupplierID ] , [ t0 ] . [ CategoryID ] FROM [ dbo ] . [ Products ] AS [ t0 ] GROUP BY [ t0 ] . [ CategoryID ] , [ t0 ] . [ SupplierID ] -- Context: SqlProvider(Sql2005) Model: AttributedMetaModel Build: 2.0.20612.0 先按CategoryID，再按SupplierID ，和匿名类中的循序一样。
最后一个例子。 var categories = from p in db.Products group p by new { Criterion = p.UnitPrice > 10 } into g select g; 按产品单价是否大于１０分类。其结果为两类，大于的是一类，小于及等于为另一类。好了，剩下的，大家自己多去领会。
TrackBack:http://www.cnblogs.com/126/archive/2007/07/10/812621.html