如何使用动态列透视正常的sql查询

时间:2012-07-08 00:30:48

标签: c# sql-server linq pivot

我有一个Sql查询,返回如下所示:

Rating as at    Rating type    Rating
--------------------------------------
6/7/2012        Type1          A
6/7/2012        Type2          A+
6/7/2012        Type3          B
8/7/2012        Type1          C
8/7/2012        Type2          C+
8/7/2012        Type3          B
8/7/2012        Type4          A

正如您所看到的,评级类型是动态的,我想在枢轴中显示它,但我真的不知道如何实现这一点。我想要的最终结果如下:

Rating as at    Type1   Type2   Type3   Type4
6/7/2012        A       A+       B
8/7/2012        C       C       C+       A

我想知道如何使用sql实现这一目标。或者我最好如何使用LINQ C#??

帮助将不胜感激。感谢。

3 个答案:

答案 0 :(得分:1)

select "Rating as at", 
max(case when Rating_Type = 'Type1' then Rating else 0 end) as Type1,
max(case when Rating_Type = 'Type2' then Rating else 0 end) as Type2,
max(case when Rating_Type = 'Type3' then Rating else 0 end) as Type3,
max(case when Rating_Type = 'Type4' then Rating else 0 end) as Type4,
from Table
group by "Rating as at"

答案 1 :(得分:1)

根据以下OP评论更新了答案 - 此更新现在将结果转换为字母成绩

以下使用SQL Server的Pivot运算符(MSDN)和EXEC语句(MSDN)的组合。

T-SQL解决方案将处理完全动态的列。

--Create a temp table to hold values
CREATE TABLE #MyTable (
[Rating as at] DATE,
[Rating type] Nvarchar(30),
Rating FLOAT )

INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('6/7/2012','Type1', 1)
INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('6/7/2012','Type2', 3)
INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('6/7/2012','Type3', 5)
INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('8/7/2012','Type1', 5)
INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('8/7/2012','Type2', 2)
INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('8/7/2012','Type3', 4)
INSERT INTO #MyTable ([Rating as at],[Rating type],Rating) VALUES ('8/7/2012','Type4', 1)
SELECT DISTINCT [Rating type] INTO #MyTableDistinct FROM #MyTable

--Create a string of dynamic fields
DECLARE @MyVar1 nvarchar(max),@MyVar2 nvarchar(max)
SELECT @MyVar1 = COALESCE(@MyVar1,'') + '[' + [Rating type] + ']' + ', '  FROM #MyTableDistinct 
SELECT @MyVar1 = LEFT(@MyVar1, LEN(@MyVar1) - 1)

--Create a string of dynamic CASE statements to be used to convert pivoted results to letter grades
--Update the CASE steatement to handle different grade types
SELECT @MyVar2 = COALESCE(@MyVar2,'') + 'CASE WHEN [' + [Rating type] + '] IS NULL THEN ''N/A'' WHEN [' + [Rating type] + '] > 4 THEN ''A'' WHEN [' + [Rating type] + '] > 3 THEN ''B'' WHEN [' + [Rating type] + '] > 2 THEN ''C'' ELSE ''F'' END AS [' + [Rating type] + '], '  FROM #MyTableDistinct 
SELECT @MyVar2 = LEFT(@MyVar2, LEN(@MyVar2) - 1)

--Build a SQL string to be later execute
--This is where all of the PIVOT magic happens
DECLARE @MySQLStatement nvarchar(max)
SET @MySQLStatement = 'SELECT [Rating as at],' + @MyVar1 +  ' INTO #MyPivotTable FROM 
    (SELECT [Rating as at],[Rating type],Rating from #MyTable) AS p1
    PIVOT (
        avg(Rating) FOR [Rating type] IN (' + @MyVar1 + ') 
    ) as p2;SELECT [Rating as at], ' + @MyVar2 + ' FROM #MyPivotTable;DROP TABLE #MyPivotTable;'

--Execute the SQL string
EXEC(@MySQLStatement)


DROP TABLE #MyTableDistinct
DROP TABLE #MyTable

答案 2 :(得分:0)

获取数据可能最简单,并使用linq进行转换。你永远不可能真正得到一个具有Type1,Type2等属性的对象,因为你需要在运行时执行此操作,但是你可以获得一组值。您首先需要获取所有评级类型的列表,然后再进行第二次查询以转动数据。像这样:

public class SomeClass
{
    public DateTime RatingDate;
    public string RatingType;
    public int Rating;
}

var data = new SomeClass[]
{
    new SomeClass() { RatingDate = DateTime.Parse("6/7/2012"), RatingType = "Type1", Rating = 1 },
    new SomeClass() { RatingDate = DateTime.Parse("6/7/2012"), RatingType = "Type2", Rating = 3 },
    new SomeClass() { RatingDate = DateTime.Parse("6/7/2012"), RatingType = "Type3", Rating = 5 },
    new SomeClass() { RatingDate = DateTime.Parse("8/7/2012"), RatingType = "Type1", Rating = 5 },
    new SomeClass() { RatingDate = DateTime.Parse("8/7/2012"), RatingType = "Type2", Rating = 2 },
    new SomeClass() { RatingDate = DateTime.Parse("8/7/2012"), RatingType = "Type3", Rating = 4 },
    new SomeClass() { RatingDate = DateTime.Parse("8/7/2012"), RatingType = "Type4", Rating = 1 }
};

var ratingTypes = data.Select(i => i.RatingType)
    .Distinct().OrderBy(i => i).ToArray();

var results = data.GroupBy(i => i.RatingDate)
    .Select(g => new { RatingDate = g.Key, Ratings = ratingTypes.GroupJoin(g, o => o, i => i.RatingType, (o, i) => i.Select(x => x.Rating).Sum()).ToArray() })
    .ToArray();

如果您愿意,可以将组连接分成单独的函数以使其更具可读性:

Func<IEnumerable<SomeClass>, int[]> func
    = group => ratingTypes.GroupJoin(group, o => o, i => i.RatingType, (o, i) => i.Select(x => x.Rating).Sum()).ToArray();

var results = data.GroupBy(i => i.RatingDate)
    .Select(g => new { RatingDate = g.Key, Ratings = func(g) })
    .ToArray();
相关问题