SQLserver中cube：多维数据集实例详解

数据库  /  管理员 发布于 5年前   267

1、cube:生成多维数据集，包含各维度可能组合的交叉表格，使用with 关键字连接 with cube

根据需要使用union all 拼接

判断 某一列的null值来自源数据还是 cube 使用GROUPING关键字

GROUPING([档案号]) = 1 ： null值来自cube（代表所有的档案号）
GROUPING([档案号]) = 0 ： null值来自源数据

举例：

SELECT * INTO ##GETFROM   (SELECT *    FROM ( SELECT      CASE      WHEN (GROUPING([档案号]) = 1) THEN      '合计'      ELSE [档案号]      END AS '档案号',      CASE      WHEN (GROUPING([系列]) = 1) THEN      '合计'      ELSE [系列]      END AS '系列',      CASE      WHEN (GROUPING([店长]) = 1) THEN      '合计'      ELSE [店长]      END AS '店长', SUM (剩余次数) AS '总剩余',      CASE      WHEN (GROUPING([店名]) = 1) THEN      '合计'      ELSE [店名]      END AS '店名'    FROM ##PudianCard    GROUP BY [档案号], [店名], [店长], [系列]    WITH cube    HAVING GROUPING([店名]) != 1        AND GROUPING([档案号]) = 1 --AND GROUPING([系列]) = 1 ) AS M    UNION    ALL       (SELECT *        FROM ( SELECT          CASE          WHEN (GROUPING([档案号]) = 1) THEN          '合计'          ELSE [档案号]          END AS '档案号',          CASE          WHEN (GROUPING([系列]) = 1) THEN          '合计'          ELSE [系列]          END AS '系列',          CASE          WHEN (GROUPING([店长]) = 1) THEN          '合计'          ELSE [店长]          END AS '店长', SUM (剩余次数) AS '总剩余',          CASE          WHEN (GROUPING([店名]) = 1) THEN          '合计'          ELSE [店名]          END AS '店名'        FROM ##PudianCard        GROUP BY [档案号], [店名], [店长], [系列]        WITH cube        HAVING GROUPING([店名]) != 1AND GROUPING([店长]) != 1 ) AS P )        UNION        ALL           (SELECT *FROM ( SELECT  CASE  WHEN (GROUPING([档案号]) = 1) THEN  '合计'  ELSE [档案号]  END AS '档案号',  CASE  WHEN (GROUPING([系列]) = 1) THEN  '合计'  ELSE [系列]  END AS '系列',  CASE  WHEN (GROUPING([店长]) = 1) THEN  '合计'  ELSE [店长]  END AS '店长', SUM (剩余次数) AS '总剩余',  CASE  WHEN (GROUPING([店名]) = 1) THEN  '合计'  ELSE [店名]  END AS '店名'FROM ##PudianCardGROUP BY [档案号], [店名], [店长], [系列]WITH cubeHAVING GROUPING([店名]) != 1    AND GROUPING([店长]) != 1 ) AS W )UNIONALL   (SELECT *    FROM ( SELECT      CASE      WHEN (GROUPING([档案号]) = 1) THEN      '合计'      ELSE [档案号]      END AS '档案号',      CASE      WHEN (GROUPING([系列]) = 1) THEN      '合计'      ELSE [系列]      END AS '系列',      CASE      WHEN (GROUPING([店长]) = 1) THEN      '合计'      ELSE [店长]      END AS '店长', SUM (剩余次数) AS '总剩余',      CASE      WHEN (GROUPING([店名]) = 1) THEN      '合计'      ELSE [店名]      END AS '店名'    FROM ##PudianCard    GROUP BY [档案号], [店名], [店长], [系列]    WITH cube    HAVING GROUPING([店名]) = 1        AND GROUPING([店长]) = 1        AND GROUPING([档案号]) = 1 ) AS K ) ) AS T

2、rollup:功能跟cube相似

3、将某一列的数据作为列名，动态加载，使用存储过程，拼接字符串

DECLARE @st nvarchar (MAX) = '';SELECT @st =@st + 'max(case when [系列]=''' + CAST ([系列] AS VARCHAR) + ''' then [总剩余] else null end ) as [' + CAST ([系列] AS VARCHAR) + '],'FROM ##GETGROUP BY [系列]; print @st;

4、根据某一列分组，分别建表

SELECT'select ROW_NUMBER() over(order by [卡项] desc) as [序号], [会员],[档案号],[卡项],[剩余次数],[员工],[店名] into ' + ltrim([店名]) + ' from 查询 where [店名]=''' + [店名] + ''' ORDER BY [卡项] desc'FROM查询GROUP BY[店名]

总结

以上就是本文关于SQLserver中cube：多维数据集实例详解的全部内容，希望对大家有所帮助。感兴趣的朋友可以继续参阅：MYSQL子查询和嵌套查询优化实例解析、几个比较重要的MySQL变量、ORACLE SQL语句优化技术要点解析等，有什么问题可以随时留言，小编会及时回复大家的。感谢各位对本站的支持！