Converting SQL Rank() to LINQ, or alternative

问题:

I have the below SQL statement that works as desired/expected. However I would like to translate it into a LINQ statement(Lambda??) so that it will fit with the rest of my DAL. However I cannot see to figure out how to simulate Rank() in LINQ.

The reason I posted it here, which is maybe in error, is to see if anyone has an alternative to the Rank() statement so that I can get this switched over. Alternatively, if there is a way to represent Rank() in LINQ that would be appreciated also.

USE CMO

SELECT      vp.[PersonID] AS [PersonId]
            ,ce.[EnrollmentID]
            ,vp.[FirstName]
            ,vp.[LastName]
            ,ce.[EnrollmentDate]
            ,ce.[DisenrollmentDate]
            ,wh.WorkerCategory

FROM  [dbo].[vwPersonInfo] AS vp
            INNER JOIN 
            (
                  [dbo].[tblCMOEnrollment] AS ce
                  LEFT OUTER JOIN
                        (
                              SELECT   *
                                          ,RANK()OVER(PARTITION BY EnrollmentID ORDER BY CASE WHEN EndDate IS NULL THEN 1 ELSE 2 END, EndDate DESC, StartDate DESC) AS whrank 
                              FROM  [dbo].[tblWorkerHistory]
                              WHERE WorkerCategory = 2
                        ) AS wh 
                              ON ce.[EnrollmentID] = wh.[EnrollmentID] AND wh.whrank = 1
            ) 
                  ON vp.[PersonID] = ce.[ClientID]

WHERE (vp.LastName NOT IN ('Client','Orientation','Real','Training','Matrix','Second','Not'))
AND (
            (wh.[EndDate] <= GETDATE())
            OR wh.WorkerCategory IS NULL
      ) 
AND (
            (ce.[DisenrollmentDate] IS NULL) 
            OR (ce.[DisenrollmentDate] >= GetDate())
      )

回答1:

Here's a sample that shows how I would simulate Rank() in Linq:

   var q = from s in class.student
                 orderby s.Age descending
                 select new { 
                     Name = s.name, 
                     Rank = (from o in class.student
                             where o.mark > s.mark
                             select o).Count() + 1 
                 };

回答2:

LINQ has rank funcionality built in, but not in the query syntax. When using the method syntax most linq functions come in two versions - the normal one and one with a rank supplied.

A simple example selecting only every other student and then adding the index in the resulting sequence to the result:

var q = class.student.OrderBy(s => s.studentId).Where((s, i) => i % 2 == 0)
.Select((s,i) => new
{
  Name = s.Name,
  Rank = i
}

回答3:

If you want to simulate rank then you can use following linq query.

      var q = (from s in class.student
                select new
                {
                    Name = s.Name,
                    Rank = (from o in class.student
                            where o.Mark > s.Mark && o.studentId == s.studentId
                            select o.Mark).Distinct().Count() + 1
                }).ToList();

you can use order by like:

      var q = (from s in class.student
                orderby s.studentId 
                select new
                {
                    Name = s.Name,
                    Rank = (from o in class.student
                            where o.Mark > s.Mark && o.studentId == s.studentId
                            select o.Mark).Distinct().Count() + 1
                }).ToList();

but order by does not matter in this query.

回答4:

Based on answer from @Totero but with a lamda implementation. Higher score = higher rank.

var rankedData = data.Select(s => new{
                    Ranking = data.Count(x => x.Value > s.Value)+1,
                    Name = s.Key,
                    Score = s.Value});

For this input:

{ 100, 100, 98, 97, 97, 97, 91, 50 }

You will get this output:

• Score : Rank
• 100 : 1
• 100 : 1
• 98 : 3
• 97 : 4
• 97 : 4
• 97 : 4
• 91 : 6
• 50 : 7