标签：
数据库环境：SQL SERVER 2008R2
有一“坐标”表t，表结构如下：
id int，
num int
字段id是序号，递增且连续，字段num是数值类型。id可以看成是坐标轴的横轴，num则跟纵轴有关系，
连续的2行记录，如果后一行的num值比前一行的num值大，则是递增趋势，反之，是递减趋势。要实现的效果如下图1：
实现思路：
将id=1的“坐标”所在纵轴设定为0，然后遍历后续所有的“坐标”，后面“坐标”的num值比前一个“坐标”的num大的，则纵坐标+1，
否则纵坐标-1。再对整理后的结果集进行行转列，根据纵坐标分组。
1.建表，插入测试数据
CREATE TABLE t(id INT,num INT) INSERT INTO t VALUES(1,1); INSERT INTO t VALUES(2,3); INSERT INTO t VALUES(3,4); INSERT INTO t VALUES(4,7); INSERT INTO t VALUES(5,5); INSERT INTO t VALUES(6,2); INSERT INTO t VALUES(7,6); INSERT INTO t VALUES(8,8); INSERT INTO t VALUES(9,4); INSERT INTO t VALUES(10,0); INSERT INTO t VALUES(11,9); INSERT INTO t VALUES(12,10); INSERT INTO t VALUES(13,12); INSERT INTO t VALUES(14,11); INSERT INTO t VALUES(15,17); INSERT INTO t VALUES(16,4); INSERT INTO t VALUES(17,2); INSERT INTO t VALUES(18,1);
2.遍历所有“坐标”，生成分组依据
WITH x1 ( id, num, gp ) AS ( SELECT id , num , 0 AS gp FROM t x0 WHERE id = 1 UNION ALL SELECT x0.id , x0.num , CASE WHEN x0.num > x1.num THEN x1.gp - 1 ELSE x1.gp + 1 END AS gp FROM t x0 , x1 WHERE x0.id = x1.id + 1 )
3.行转列实现最终结果集
SELECT ISNULL(CAST([1] AS VARCHAR(2)), ‘‘) [1] , ISNULL(CAST([2] AS VARCHAR(2)), ‘‘) [2] , ISNULL(CAST([3] AS VARCHAR(2)), ‘‘) [3] , ISNULL(CAST([4] AS VARCHAR(2)), ‘‘) [4] , ISNULL(CAST([5] AS VARCHAR(2)), ‘‘) [5] , ISNULL(CAST([6] AS VARCHAR(2)), ‘‘) [6] , ISNULL(CAST([7] AS VARCHAR(2)), ‘‘) [7] , ISNULL(CAST([8] AS VARCHAR(2)), ‘‘) [8] , ISNULL(CAST([9] AS VARCHAR(2)), ‘‘) [9] , ISNULL(CAST([10] AS VARCHAR(2)), ‘‘) [10] , ISNULL(CAST([11] AS VARCHAR(2)), ‘‘) [11] , ISNULL(CAST([12] AS VARCHAR(2)), ‘‘) [12] , ISNULL(CAST([13] AS VARCHAR(2)), ‘‘) [13] , ISNULL(CAST([14] AS VARCHAR(2)), ‘‘) [14] , ISNULL(CAST([15] AS VARCHAR(2)), ‘‘) [15] , ISNULL(CAST([16] AS VARCHAR(2)), ‘‘) [16] , ISNULL(CAST([17] AS VARCHAR(2)), ‘‘) [17] , ISNULL(CAST([18] AS VARCHAR(2)), ‘‘) [18] FROM ( SELECT * FROM x1 ) AS t1 PIVOT( MAX(num) FOR id IN ( [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18] ) )AS t2
如果要实现比较逼真的趋势图，用“/”、“\”替代数值，实现下图2的效果，只需在遍历坐标的时候再做些处理即可。
实现的SQL脚本：
WITH x1 ( id, num, cc, gp ) AS ( SELECT id , num , ‘/‘ AS cc , 0 AS gp FROM t x0 WHERE id = 1 UNION ALL SELECT x0.id , x0.num , CASE WHEN x0.num > x1.num THEN ‘/‘ ELSE ‘\‘ END AS cc , CASE WHEN x0.num > x1.num THEN x1.gp - 1 ELSE x1.gp + 1 END AS gp FROM t x0 , x1 WHERE x0.id = x1.id + 1 ) SELECT ISNULL([1], ‘‘) [1] , ISNULL([2], ‘‘) [2] , ISNULL([3], ‘‘) [3] , ISNULL([4], ‘‘) [4] , ISNULL([5], ‘‘) [5] , ISNULL([6], ‘‘) [6] , ISNULL([7], ‘‘) [7] , ISNULL([8], ‘‘) [8] , ISNULL([9], ‘‘) [9] , ISNULL([10], ‘‘) [10] , ISNULL([11], ‘‘) [11] , ISNULL([12], ‘‘) [12] , ISNULL([13], ‘‘) [13] , ISNULL([14], ‘‘) [14] , ISNULL([15], ‘‘) [15] , ISNULL([16], ‘‘) [16] , ISNULL([17], ‘‘) [17] , ISNULL([18], ‘‘) [18] FROM ( SELECT id , cc , gp FROM x1 ) AS t1 PIVOT( MAX(cc) FOR id IN ( [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18] ) )AS t2
标签：
原文地址：http://www.cnblogs.com/boss-he/p/4758151.html
