1.rollup:rollup 是group by 的扩展，它只能出现在group by后面。主要是为了解决多层分组的问题。语法：
SELECT . . .
FROM . . .
GROUP BY ROLLUP (ordered list of grouping columns)
例如要查询每个部门的工资总和所有部门的总和
TEST@orcl> select dept_id,sum(salary) from employee group by rollup(dept_id);
DEPT_ID SUM(SALARY)
---------- -----------
10 8750
20 9900
30 9400
<null> 28050
如果rollup()有两列，将会有3个分组例如：
group by rollup(year，month)相当于
group by year,month
union
group by year
union
group by null;
----------------------------------------------
TEST@orcl> select year,month,sum(tot_sales) from all_orders
2 group by rollup(year,month);
YEAR MONTH SUM(TOT_SALES)
---------- ---------- --------------
2000 1 2997866
2000 2 3325690
2000 3 3719014
2000 4 3497570
2000 5 3640628
2000 6 3146822
2000 7 3446480
2000 8 2973882
2000 9 3207396
2000 10 3193212
2000 11 3096300
2000 12 3278944
2000 <null> 39523804
2001 1 1498933
2001 2 1662845
2001 3 1859507
2001 4 1748785
2001 5 1820314
2001 6 1573411
2001 7 1723240
2001 8 1486941
2001 9 1603698
2001 10 1596606
2001 11 1548150
2001 12 1639472
2001 <null> 19761902
<null> <null> 59285706
1.1 部分rollup：group by + 分组列 rollup(分组列表)
group by year ,rollup (month) 相当于
group by year ,month
union
group year;
1 select year,month,sum(tot_sales) from all_orders
2* group by year, rollup(month)
TEST@orcl> /
YEAR MONTH SUM(TOT_SALES)
---------- ---------- --------------
2000 1 2997866
2000 2 3325690
2000 3 3719014
2000 4 3497570
2000 5 3640628
2000 6 3146822
2000 7 3446480
2000 8 2973882
2000 9 3207396
2000 10 3193212
2000 11 3096300
2000 12 3278944
2000 <null> 39523804
2001 1 1498933
2001 2 1662845
2001 3 1859507
2001 4 1748785
2001 5 1820314
2001 6 1573411
2001 7 1723240
2001 8 1486941
2001 9 1603698
2001 10 1596606
2001 11 1548150
2001 12 1639472
2001 <null> 19761902
2.CUBE:cube类似于rollup,但产生的分组是分组类表的排列组合
SELECT . . .
FROM . . .
GROUP BY CUBE (list of grouping columns)
cube(year,month) 相当于
group by year,month
union
group by year
union
group by month
unin
group by null;
例子：
1 select year,month,sum(tot_sales) from all_orders
2* group by cube(year,month)
TEST@orcl> /
YEAR MONTH SUM(TOT_SALES)
---------- ---------- --------------
<null> <null> 59285706
<null> 1 4496799
<null> 2 4988535
<null> 3 5578521
<null> 4 5246355
<null> 5 5460942
2.1 部分CUBE group by 分组列,cube(分组列表)
group by 后面的分组列可以看做一个固定列
3.GROUPING 函数，rollup和cube分组输出的结果是多级分组的,grouping函数能区分分组的级别。例如：
1 select year,month,grouping(year) g_year,grouping(month) g_month from all_orders
2* group by rollup(year,month)
TEST@orcl> /
YEAR MONTH G_YEAR G_MONTH
---------- ---------- ---------- ----------
2000 1 0 0
2000 2 0 0
2000 3 0 0
2000 4 0 0
2000 5 0 0
2000 6 0 0
2000 7 0 0
2000 8 0 0
2000 9 0 0
2000 10 0 0
2000 11 0 0
2000 12 0 0
2000 <null> 0 1
2001 1 0 0
2001 2 0 0
2001 3 0 0
2001 4 0 0
2001 5 0 0
2001 6 0 0
2001 7 0 0
2001 8 0 0
2001 9 0 0
2001 10 0 0
2001 11 0 0
2001 12 0 0
2001 <null> 0 1
<null> <null> 1 1
grouping(分组列) 返回0或1，0:代表该分组列参加分组 1:不参加分组，grouping主要用在格式化输出 例如：
1 select decode(grouping(year),1,'整年',year) year,decode(grouping(month),1,'整月',month) month
2 from all_orders
3* group by rollup(year,month)
TEST@orcl> /
YEAR MONTH
---------------------------------------- ----------------------------------------
2000 1
2000 2
2000 3
2000 4
2000 5
2000 6
2000 7
2000 8
2000 9
2000 10
2000 11
2000 12
2000 整月
2001 1
2001 2
2001 3
2001 4
2001 5
2001 6
2001 7
2001 8
2001 9
2001 10
2001 11
2001 12
2001 整月
整年 整月
4.grouping sets:类似于rollup和cube
grouping sets(year,monh) 相当于
group by year
union
group by month
例如：
1 select decode(grouping(year),1,'整年',year) year,decode(grouping(month),1,'整月',month) month
2 from all_orders
3* group by grouping sets(year,month)
TEST@orcl> /
YEAR MONTH
---------------------------------------- ----------------------------------------
整年 1
整年 11
整年 6
整年 2
整年 4
整年 5
整年 8
整年 3
整年 7
整年 10
整年 12
整年 9
2001 整月
2000 整月
5.GROUP_ID和GROUPING_ID函数
5.1 GROUPING_ID函数：
语法：
SELECT . . . , GROUPING_ID(ordered_list_of_grouping_columns)
FROM . . .
GROUP BY . . .
grouping_id函数和grouping函数相关联的,grouping函数返回的是0或者1,grouping_id 返回的就是对应列的grouping的值然后再组合，最后转换成十进制。例如：
1 select year,month,grouping(year) g_year,grouping(month) g_month from all_orders
2 where year=2000
3* group by rollup(year,month)
SQL> /
YEAR MONTH G_YEAR G_MONTH
---------- ---------- ---------- ----------
2000 1 0 0
2000 2 0 0
2000 3 0 0
2000 4 0 0
2000 5 0 0
2000 6 0 0
2000 7 0 0
2000 8 0 0
2000 9 0 0
2000 10 0 0
2000 11 0 0
YEAR MONTH G_YEAR G_MONTH
---------- ---------- ---------- ----------
2000 12 0 0
2000 0 1
1 1
grouping_id(year,month)就是将G_YEAR和G_MONTH的值用||连接起来，因为grouping返回的0或1，可以看出是2进制，最终的结果会以十进制展示，例如：
1 select year,month,grouping(year) g_year,grouping(month) g_month,grouping_id(year,month) g_y_m,
2 grouping_id(month,year) g_m_y from all_orders
3 where year=2000
4* group by rollup(year,month)
SQL> /
YEAR MONTH G_YEAR G_MONTH G_Y_M G_M_Y
---------- ---------- ---------- ---------- ---------- ----------
2000 1 0 0 0 0
2000 2 0 0 0 0
2000 3 0 0 0 0
2000 4 0 0 0 0
2000 5 0 0 0 0
2000 6 0 0 0 0
2000 7 0 0 0 0
2000 8 0 0 0 0
2000 9 0 0 0 0
2000 10 0 0 0 0
2000 11 0 0 0 0
YEAR MONTH G_YEAR G_MONTH G_Y_M G_M_Y
---------- ---------- ---------- ---------- ---------- ----------
2000 12 0 0 0 0
2000 0 1 1 2
1 1 3 3
其作用可以用来过滤返回的分组级别。
5.2 GROUP_ID
语法：
SELECT . . . , GROUP_ID( )
FROM . . .
GROUP BY . . .
group_id 分组可以识别哪些是重复的分组,唯一的分组返回0，例如：
1 select year,month,grouping(year) g_year,grouping(month) g_month,group_id() g from all_orders
2 where year=2000
3* group by year ,rollup(year,month)
QL> /
YEAR MONTH G_YEAR G_MONTH G
--------- ---------- ---------- ---------- ----------
2000 1 0 0 0
2000 2 0 0 0
2000 3 0 0 0
2000 4 0 0 0
2000 5 0 0 0
2000 6 0 0 0
2000 7 0 0 0
2000 8 0 0 0
2000 9 0 0 0
2000 10 0 0 0
2000 11 0 0 0
YEAR MONTH G_YEAR G_MONTH G
--------- ---------- ---------- ---------- ----------
2000 12 0 0 0
2000 0 1 0
2000 0 1 1
可以用 HAVING GROUP_ID() < 1过滤掉重复的分组。