GROUP BY的扩展
GROUP BY的扩展主要包括ROLLUP，CUBE，GROUPING SETS三种形式。
ROLLUP
rollup相对于简单的分组合计增加了小计和合计，解释起来会比较抽象，下面我们来看看具体事例。
例1，统计不同部门工资的总和和所有部门工资的总和。
SQL> select deptno,sum(sal) from emp group by rollup(deptno);
DEPTNO SUM(SAL)
---------- ----------
10 8750
20 10875
30 9400
29025
例2，该例中先对deptno进行分组，再对job进行分组
SQL> select deptno,job,sum(sal) from emp group by rollup(deptno,job);
DEPTNO JOB SUM(SAL)
---------- --------- ----------
10 CLERK 1300 --10号部门中JOB为CLERK的工资的总和
10 MANAGER 2450
10 PRESIDENT 5000
10 8750 --10号所有工种工资的总和
20 CLERK 1900
20 ANALYST 6000
20 MANAGER 2975
20 10875
30 CLERK 950
30 MANAGER 2850
30 SALESMAN 5600
30 9400
29025 --所有部门，所有工种工资的总和
13 rows selected.
如果要用普通的分组函数实现，可用UNION ALL语句：
--实现单个部门，单个工种的工资的总和
select deptno,job,sum(sal) from emp group by deptno,job
union all
--实现单个部门工资的总和
select deptno,null,sum(sal) from emp group by deptno
union all
--实现所有部门工资的总和
select null,null,sum(sal) from emp
order by 1,2
下面我们分别来看看两者的执行计划及统计信息，
ROLLUP语句：
Execution Plan
-----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 11 | 132 | 3 (34)| 00:00:01 |
| 1 | SORT GROUP BY ROLLUP| | 11 | 132 | 3 (34)| 00:00:01 |
| 2 | TABLE ACCESS FULL | EMP | 14 | 168 | 2 (0)| 00:00:01 |
-----------------------------------------------------------------------------
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
2 consistent gets
0 physical reads
0 redo size
895 bytes sent via SQL*Net to client
519 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
13 rows processed
UNION ALL语句：
Execution Plan
-----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 15 | 150 | 9 (34)| 00:00:01 |
| 1 | SORT ORDER BY | | 15 | 150 | 8 (75)| 00:00:01 |
| 2 | UNION-ALL | | | | | |
| 3 | HASH GROUP BY | | 11 | 132 | 3 (34)| 00:00:01 |
| 4 | TABLE ACCESS FULL| EMP | 14 | 168 | 2 (0)| 00:00:01 |
| 5 | HASH GROUP BY | | 3 | 15 | 3 (34)| 00:00:01 |
| 6 | TABLE ACCESS FULL| EMP | 14 | 70 | 2 (0)| 00:00:01 |
| 7 | SORT AGGREGATE | | 1 | 3 | | |
| 8 | TABLE ACCESS FULL| EMP | 14 | 42 | 2 (0)| 00:00:01 |
-----------------------------------------------------------------------------
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
6 consistent gets
0 physical reads
0 redo size
895 bytes sent via SQL*Net to client
519 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
13 rows processed
不难看出，相同的功能实现，ROLLUP相对于UNION ALL效率有了极大的提升。
CUBE
cube相对于rollup，结果输出更加详细。
例1，在本例中还不是很明显。
SQL> select deptno,sum(sal) from emp group by cube(deptno);
DEPTNO SUM(SAL)
---------- ----------
29025
10 8750
20 10875
30 9400
例2，相对于rollup，cube还对工种这一列进行了专门的汇总。
SQL> select deptno,job,sum(sal) from emp group by cube(deptno,job);
DEPTNO JOB SUM(SAL)
---------- --------- ----------
29025
CLERK 4150
ANALYST 6000
MANAGER 8275
SALESMAN 5600
PRESIDENT 5000
10 8750
10 CLERK 1300
10 MANAGER 2450
10 PRESIDENT 5000
20 10875
20 CLERK 1900
20 ANALYST 6000
20 MANAGER 2975
30 9400
30 CLERK 950
30 MANAGER 2850
30 SALESMAN 5600
18 rows selected.
GROUPING SETS
GROUPING SETS相对于ROLLUP和CUBE，结果是分类统计的，可读性更好一些。
例1：
SQL> select deptno,job,to_char(hiredate,'yyyy')hireyear,sum(sal) from emp group by grouping sets(deptno,job,to_char(hiredate,'yyyy'));
DEPTNO JOB HIRE SUM(SAL)
---------- --------- ---- ----------
CLERK 4150
SALESMAN 5600
PRESIDENT 5000
MANAGER 8275
ANALYST 6000
30 9400
20 10875
10 8750
1987 4100
1980 800
1982 1300
1981 22825
例2：
SQL> select deptno,job,sum(sal) from emp group by grouping sets(deptno,job);
DEPTNO JOB SUM(SAL)
---------- --------- ----------
CLERK 4150
SALESMAN 5600
PRESIDENT 5000
MANAGER 8275
ANALYST 6000
30 9400
20 10875
10 8750
8 rows selected.
对于该例，如何用UNION ALL实现呢？
select null deptno,job,sum(sal) from emp group by job
union all
select deptno,null,sum(sal) from emp group by deptno;
两者的执行计划及统计信息分别如下：
GROUPING SETS：
Execution Plan
--------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 11 | 352 | 10 (20)| 00:00:01 |
| 1 | TEMP TABLE TRANSFORMATION | | | | | |
| 2 | LOAD AS SELECT | SYS_TEMP_0FD9D6795_E71F79 | | | | |
| 3 | TABLE ACCESS FULL | EMP | 14 | 168 | 2 (0)| 00:00:01 |
| 4 | LOAD AS SELECT | SYS_TEMP_0FD9D6796_E71F79 | | | | |
| 5 | HASH GROUP BY | | 1 | 19 | 3 (34)| 00:00:01 |
| 6 | TABLE ACCESS FULL | SYS_TEMP_0FD9D6795_E71F79 | 1 | 19 | 2 (0)| 00:00:01 |
| 7 | LOAD AS SELECT | SYS_TEMP_0FD9D6796_E71F79 | | | | |
| 8 | HASH GROUP BY | | 1 | 26 | 3 (34)| 00:00:01 |
| 9 | TABLE ACCESS FULL | SYS_TEMP_0FD9D6795_E71F79 | 1 | 26 | 2 (0)| 00:00:01 |
| 10 | VIEW | | 1 | 32 | 2 (0)| 00:00:01 |
| 11 | TABLE ACCESS FULL | SYS_TEMP_0FD9D6796_E71F79 | 1 | 32 | 2 (0)| 00:00:01 |
--------------------------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
4 recursive calls
24 db block gets
17 consistent gets
3 physical reads
1596 redo size
819 bytes sent via SQL*Net to client
519 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
8 rows processed
UNION ALL：
----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 8 | 65 | 6 (67)| 00:00:01 |
| 1 | UNION-ALL | | | | | |
| 2 | HASH GROUP BY | | 5 | 50 | 3 (34)| 00:00:01 |
| 3 | TABLE ACCESS FULL| EMP | 14 | 140 | 2 (0)| 00:00:01 |
| 4 | HASH GROUP BY | | 3 | 15 | 3 (34)| 00:00:01 |
| 5 | TABLE ACCESS FULL| EMP | 14 | 70 | 2 (0)| 00:00:01 |
---------------------------------------------------------------------------- Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
4 consistent gets
0 physical reads
0 redo size
819 bytes sent via SQL*Net to client
519 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
8 rows processed
和rollup不同的是，grouping sets的效率竟然比同等功能的union all语句低，这实现有点出乎意料。看来，也不可盲目应用Oracle提供的方案，至少，在本例中是如此。
GROUP BY的扩展的更多相关文章
- 报表开发之扩展GROUP BY
在实际运用中.比方在数据仓库中,常常须要对数据进行多维分析.不仅须要标准分组的结果(相当于 GROUP BY),还须要不同维度的小计(简单 GROUP BY 中取部分列分组)和合计(不分组).从而 提 ...
- mysql5.5手册读书日记(3)
<?php /* MySQL_5.5中文参考手册 587开始 与GROUP BY子句同时使用的函数和修改程序 12.10.1. GROUP BY(聚合)函数 12.10.2. GROUP BY修 ...
- oracle 高级分组
oracle 高级分组 博客分类: 数据库基础 oraclesql 10.高级分组 本章目标: 对于增强的group by需要掌握: 1.使用rollup(也就是roll up累计的意思)操作产生s ...
- spark2.2 DataFrame的一些算子操作
Spark Session中的DataFrame类似于一张关系型数据表.在关系型数据库中对单表或进行的查询操作,在DataFrame中都可以通过调用其API接口来实现.可以参考,Scala提供的Dat ...
- Spark-SQL之DataFrame操作大全
Spark SQL中的DataFrame类似于一张关系型数据表.在关系型数据库中对单表或进行的查询操作,在DataFrame中都可以通过调用其API接口来实现.可以参考,Scala提供的DataFra ...
- python新手菜鸟之基础篇
s=0 for i in range(1,101): s += i else: print(s) def main(n): '''打印菱形图形''' for i in range(n): print( ...
- Spark-SQL之DataFrame操作
Spark SQL中的DataFrame类似于一张关系型数据表.在关系型数据库中对单表或进行的查询操作,在DataFrame中都可以通过调用其API接口来实现.可以参考,Scala提供的DataFra ...
- Linux系统编程【转】
转自:https://blog.csdn.net/majiakun1/article/details/8558308 一.Linux系统编程概论 1.1 系统编程基石 syscall: libc:标准 ...
- [转]详解Oracle高级分组函数(ROLLUP, CUBE, GROUPING SETS)
原文地址:http://blog.csdn.net/u014558001/article/details/42387929 本文主要讲解 ROLLUP, CUBE, GROUPING SETS的主要用 ...
随机推荐
- sublime Text 3 字体
1,Comic Sans Ms 2,DejaVu Sans Mono 3,microsoft yahei(微软雅黑) 4,Microsoft Yahei UI(微软雅黑增强版) 5,Ubuntu Mo ...
- 各种android应用模仿源码
V2EX的非官方Android客户端,极力遵循Material Design风格 下载 湘潭大学三翼校园"四季电台" Android客户端 下载 高仿煎蛋客户端 下载 ...
- ArcGIS10的附件功能
转自 积思园 http://blog.csdn.net/linghe301/article/details/6386176 老是忘记怎么使用这个ArcGIS10的附件功能,这次就做个记录吧. 在项目应 ...
- 关于IoCallDriver使用的疑惑
#pragma PAGEDCODE NTSTATUS HelloDDKRead(IN PDEVICE_OBJECT pDevObj, IN PIRP pIrp) { KdPrint(("Dr ...
- $Host.Runspace.ThreadOptions = “ReuseThread”有神马用？
$Host.Runspace.ThreadOptions = “ReuseThread” 在很多PowerShell的脚本中你都会看到这个语句被用来开头,那它的作用是什么呢? 答:这个设置可以提高对内 ...
- cocoapods 命令
1.使用CocoaPods a 新建一个项目,名字cocoapods
- Unity 5.3.5p8 C#编译器升级
Unity 5.3.5p8的C#编译器升级 注意:该版本是单独升级C#编译器的测试版!请使用文中提供的下载链接! 基于Unity 5.3.5p8的C#编译器升级!下载链接 试用该版本前请先备份项目,遇 ...
- 使用 OWIN Self-Host ASP.NET Web API 2
Open Web Interface for .NET (OWIN)在Web服务器和Web应用程序之间建立一个抽象层.OWIN将网页应用程序从网页服务器分离出来,然后将应用程序托管于OWIN的程序而离 ...
- .net中事件引起的内存泄漏分析
系列主题:基于消息的软件架构模型演变 在Winform和Asp.net时代,事件被大量的应用在UI和后台交互的代码中.看下面的代码: private void BindEvent() { var bt ...
- MySQL MVCC(多版本并发控制)
概述 为了提高并发MySQL加入了多版本并发控制,它把旧版本记录保存在了共享表空间(undolog),当事务提交之后将重做日志写入磁盘(前提innodb_flush_log_at_trx_commit ...