【T-SQL系列】新的排序函数

如：ROW_NUMBER、RANK、DENSE_RANK
三个分析函数都是按照col1分组内从1开始排序

ROW_NUMBER() 是没有重复值的排序(即使两天记录相等也是不重复的)，可以利用它来实现分页
DENSE_RANK() 是连续排序，两个第二名仍然跟着第三名
RANK() 是跳跃拍学，两个第二名下来就是第四名

示例：

DECLARE @t1 TABLE
    (
      Sequence INT ,
      Name VARCHAR(20) ,
      Score INT
    )

INSERT  INTO @t1
        SELECT  1 ,
                '甲' ,
                7
        UNION ALL
        SELECT  2 ,
                '甲' ,
                8
        UNION ALL
        SELECT  3 ,
                '甲' ,
                8
        UNION ALL
        SELECT  4 ,
                '甲' ,
                8
        UNION ALL
        SELECT  5 ,
                '甲' ,
                9
        UNION ALL
        SELECT  1 ,
                '乙' ,
                10
        UNION ALL
        SELECT  2 ,
                '乙' ,
                6
        UNION ALL
        SELECT  3 ,
                '乙' ,
                10
        UNION ALL
        SELECT  4 ,
                '乙' ,
                6
        UNION ALL
        SELECT  5 ,
                '乙' ,
                8

SELECT  Sequence ,
        Name ,
        Score ,
        ROW_NUMBER() OVER ( ORDER BY Sequence, Score ) AS RowNumber1 ,
        ROW_NUMBER() OVER ( ORDER BY Sequence, Score DESC ) AS RowNumber2 ,
        RANK() OVER ( ORDER BY Score ) AS Rnk ,
        DENSE_RANK() OVER ( ORDER BY Score ) AS DenseRnk ,
        NTILE(15) OVER ( ORDER BY Score ) AS Bucket
FROM    @t1

结果集：