c# - Entity Framework : multiple from is generating cross join

Question

我是 Entity Framework 的新手，我正在处理查询并找到了解决方案，但在使用 LINQPad 进行一些调查后，我不喜欢生成的 SQL。

Linq 查询是:

from p in Person
from pcc in p.PersonCreditCard
group p by new { pcc.CreditCard.CardType } into g
select new
{
     cname = g.Key.CardType,
     cptPerson = g.Count()
}

生成的sql查询是

SELECT COUNT(*) AS [cptPerson], [t2].[CardType] AS [cname]
FROM [Person].[Person] AS [t0]
CROSS JOIN [Sales].[PersonCreditCard] AS [t1]
INNER JOIN [Sales].[CreditCard] AS [t2] ON [t2].[CreditCardID] = [t1].[CreditCardID]
WHERE [t1].[BusinessEntityID] = [t0].[BusinessEntityID]
GROUP BY [t2].[CardType

注意交叉连接。据我所知，使用 CROSS JOIN 并不是一个好主意。为什么会这样？

我可以使用另一种语法来获得相同的结果(按信用卡类型计算人数)

Answer 1

尝试通过 PersonCreditCard 开始查询:

var query= context.PersonCreditCard.GroupBy(c=>c.CreditCard.CardType)
                                   .Select(g=>new {
                                                    cname = g.Key,
                                                    cptPerson = g.Count()
                                                  }
                                          );

更新

我想在 PersonCreditCard 实体中你也有一个对 Person 的引用，所以你可以像这样向你的查询添加一个 Where 调用:

var query= context.PersonCreditCard.Where(pc=>pc.Person.Name=="John")
                                   .GroupBy(c=>c.CreditCard.CardType)
                                   .Select(g=>new {
                                                    cname = g.Key,
                                                    cptPerson = g.Count()
                                                  }
                                          );

更新2

我想我明白你的意思了，如果你想在结果中包含属于每个组的人，你可以执行以下操作:

var query= context.PersonCreditCard.Where(pc=>pc.Person.Name=="John")
                                   .GroupBy(c=>c.CreditCard.CardType)
                                   .Select(g=>new {
                                                    cname = g.Key,
                                                    cptPerson = g.Count()
                                                    people=g.Select(pc=>pc.Person);
                                                  }
                                          );