jessen163
- 浏览: 457198 次
- 性别:
- 来自: 潘多拉
-
文章分类
最新评论
-
lijunwyf：
代码好像不完整，后端没串起来的感觉
Json-RPC for java首次尝试 -
jerry.chen：
我用这种方式去，貌似没啥效果，楼主！
Xfire开发的webservice客户端超时问题解决 -
漫步邃思：
这个问题常遇到，老是想不起来解决方法，记住了
resin3.1.0无法解析EL表达式问题解决 -
dampce032：
在不删掉integratedSecurity=true;的情况 ...
JDBC 连接SQLServer数据库（Failed to load the sqljdbc_auth.dll） -
kill_e680：
取CPU号和取硬盘号，在linux下可以用吗？
sigar使用：在web中应用sigar取得系统信息
使用下面的日期时间函数：
-TZ_OFFSET
-CURRENT_DATE
-CURRENT_TIMESTAMP
-LOCALTIMESTAMP
-DBTIMEZONE
-SESSIONTIMEZONE
-EXTRACT
-FROM_TZ
-TO_TIMESTAMP
-TO_TIMESTAMP_TZ
-TO_YMINTERVAL
oracle9i日期时间支持
:: 在 Oracle9i 中，你能够在你的日期和时间数据中包含时区，并且提供对小数秒的支持
:: Oracle9i 中新的日期时间数据类型：
-TIMESTAMP
-TIMESTAMP WITH TIME ZONE (TSTZ)
-TIMESTAMP WITH LOCAL TIME ZONE (TSLTZ)
:: Oracle9i 服务器提供对日期时间数据类型的夏令时支持
--CURRENT_DATE
:: 显示在会话时区中的当前日期和时间
ALTER SESSION SET NLS_DATE_FORMAT='DD-MON-YYYY HH24:MI:SS';
:: CURRENT_DATE 对于会话时区是敏感的
Note:the ALTER SESSION command sets the date format of the session to 'DD-MON-YYYY HH24:MI:SS' that is Day of month(1-31)-Abbreviated name of month -4-digit year Hours of dat(0-23):Minute(0-59):Second(0-59)
--CURRENT_TIMESTAMP
:: 显示在会话时区中的当前日期和时间,包含小数秒
ALTER SESSION SET TIME_ZONE='-5:0'; //'-8.0'
SELECT SESSIONTIMEZONE,CURRENT_TIMESTAMP FROM DUAL;
::CURRENT_TIMESTAMP 对于会话时区是敏感的
::返回值是TIMESTAMP WITH TIME ZONE 数据类型
--LOCALTIMESTAMP
:: 以 TIMESTAMP 数据类型的值显示在会话时区中的当前日期和时间
ALTER SESSION SET TIME_ZONE = '-5:0';
SELECT CURRENT_TIMESTAMP, LOCALTIMESTAMP
FROM DUAL;
:: LOCALTIMESTAMP 返回一个 TIMESTAMP 值，而 CURRENT_TIMESTAMP 返回一个 TIMESTAMP WITH TIME ZONE 值
--DBTIMEZONE 和 SESSIONTIMEZONE
:: 显示数据库时区的值
SELECT DBTIMEZONE FROM DUAL;
:: 显示会话时区的值
SELECT SESSIONTIMEZONE FROM DUAL;
--EXTRACT
:: 从SYSDATE 中显示年
SELECT EXTRACT (YEAR FROM SYSDATE) FROM DUAL;
:: 从 HIRE_DATE 中显示 MANAGER_ID 是100的雇员的月
SELECT last_name, hire_date,
EXTRACT (MONTH FROM HIRE_DATE)
FROM employees
WHERE manager_id = 100;
LAST_NAME HIRE_DATE EXTRACT(MONTHFROMHIRE_DATE)
------------------------- ---------- ---------------------------
Kochhar 21-9月 -89 9
De Haan 13-1月 -93 1
Raphaely 07-12月-94 12
Weiss 18-7月 -96 7
Fripp 10-4月 -97 4
Kaufling 01-5月 -95 5
语法:
SELECT EXTRACT ([YEAR] [MONTH][DAY] [HOUR] [MINUTE][SECOND] [TIMEZONE_HOUR] [TIMEZONE_MINUTE]
[TIMEZONE_REGION] [TIMEZONE_ABBR]
FROM [datetime_value_expression]
[interval_value_expression]);
TIMESTAMP WITH LOCAL TIME ZONE 存储在数据库时区中，当一个用户选择数据时，值被调整到用户会话的时区。
--***************增强 GROUP BY 子句*********************
目标
完成本课后, 您应当能够：
:: 用 ROLLUP 操作产生小计值
:: 用 CUBE 操作产生交叉表值
:: 用 GROUPING 函数确定由 ROLLUP 或 CUBE 创建的行值
:: 用 GROUPING SETS 产生一个单个的结果集
Lesson Aim
In this lesson you learn how to:
:: Group data for obtaining the following:
-Subtotal values by using the ROLLUP operator
-Cross-tabulation values by using the CUBE operator
:: Use the GROUPING function to identify the level of aggregation in the results set produced by a ROLLUP or CUBE operator.
:: Use GROUPING SETS to produce a single result set(结果集) that is equivalent to a UNION ALL approach(方法,步骤).
--**************组函数的回顾*********************
组函数在行集上操作，给出分组结果
SELECT [column,] group_function(column). . .
FROM table
[WHERE condition]
[GROUP BY group_by_expression]
[ORDER BY column];
例子:
SELECT AVG(salary), STDDEV(salary),
COUNT(commission_pct),MAX(hire_date)
FROM employees
WHERE job_id LIKE 'SA%';
Group Functions
You can use the GROUP BY clause to divide the rows in a table into groups. You can then use the group functions to return summary information for each group. Group functions can appear in select lists and in ORDER BY and HAVING clauses. The Oracle Server applies the group functions to each group of rows and returns a single result row for each group.
Types of Group Functions
Each of the group functions AVG, SUM, MAX, MIN, COUNT, STDDEV, and VARIANCE accept one argument. The functions AVG, SUM, STDDEV, and VARIANCE operate only on numeric values(只运算数字类型的值). MAX and MIN can operate on numeric, character, or date data values. COUNT returns the number of nonnull rows for the given expression.
The example in the slide calculates the average salary, standard deviation on the salary, number of employees earning a commission and the maximum hire date for those employees whose JOB_ID begins with SA.
Guidelines for Using Group Functions ././././.
:: The data types for the arguments can be CHAR, VARCHAR2, NUMBER, or DATE.
:: All group functions except COUNT(*) ignore null values. To substitute(为NULL替换) a value for null values, use the NVL function. COUNT returns either a number or zero.
:: The Oracle Server implicitly sorts(隐含) the results set in ascending order of the grouping columns specified(ASC), when you use a GROUP BY clause. To override this default ordering, you can use DESC in an ORDER BY clause.
Instructor Note
You can skip this slide if the students are already familiar with these concepts.
Group By 子句的回顾
语法:
SELECT [column,] group_function(column). . .
FROM table
[WHERE condition]
[GROUP BY group_by_expression]
[ORDER BY column];
Example:
SELECT department_id,job_id,SUM(salary),COUNT(employee_id)
FROM employees
GROUP BY department_id,job_id;
Review of GROUP BY Clause (回顾)
The example illustrated in the slide is evaluated by the Oracle Server as follows:
:: The SELECT clause specifies that the following columns are to be retrieved(找到):
-Department ID and job ID columns from the EMPLOYEES table
-The sum of all the salaries and the number of employees in each group that you have specified in the GROUP BY clause ../././
:: The GROUP BY clause specifies how the rows should be grouped in the table. The total salary and the number of employees are calculated for each job ID within each department. The rows are grouped by department ID and then grouped by job within each department.
HAVING 子句的回顾
SELECT [column,] group_function(column)...
FROM table
[WHERE condition]
[GROUP BY group_by_expression]
[HAVING having_expression]
[ORDER BY column];
:: 用HAVING 子句指定将被显示的那些组
:: 在限制条件的基础上可以进一步约束分组.
The HAVING Clause
Groups are formed and group functions are calculated before the HAVING clause is applied to the groups. The HAVING clause can precede the GROUP BY clause, but it is recommended that you place the GROUP BY clause first because it is more logical.
./././././
The Oracle Server performs the following steps when you use the HAVING clause:
1.Groups rows
2.Applies the group functions to the groups and displays the groups that match the criteria in the HAVING clause
SELECT department_id, AVG(salary)
FROM employees
GROUP BY department_id
HAVING AVG(salary) >9500;
DEPARTMENT_ID AVG(SALARY)
80 10033.3434
90 10036.4
100 10032.34
the example dispalys depatment ID and average salary for those departments whose average salary is greater than $9,500
GROUP BY 带 ROLLUP 或 CUBE 操作
:: 带 ROLLUP 或 CUBE 的 GROUP BY 子句用交叉引用列产生超合计行
:: ROLLUP 分组产生一个包含常规分组行和subtotal的结果集
:: CUBE 分组产生一个包含 ROLLUP 行和交叉表行的结果集
GROUP BY with the ROLLUP and CUBE Operators
You specify ROLLUP and CUBE operators in the GROUP BY clause of a query. ROLLUP grouping produces(产生) a results set containing the regular grouped rows and subtotal rows. The CUBE operation in the GROUP BY clause groups the selected rows based on the values of all possible combinations of expressions in the specification and returns a single row of summary information for each group. You can use the CUBE operator to produce cross-tabulation rows.
Note: When working with ROLLUP and CUBE, make sure that the columns following the GROUP BY clause have meaningful, real-life relationships with each other; otherwise the operators return irrelevant information(不相关的信息).
The ROLLUP and CUBE operators are available only in Oracle8i and later releases.
ROLLUP操作
SELECT [column,] group_function(column). . .
FROM table
[WHERE condition]
[GROUP BY [ROLLUP] group_by_expression]
[HAVING having_expression];
[ORDER BY column];
:: ROLLUP 是一个 GROUP BY 子句的扩展
:: 用 ROLLUP 操作产生小计和累计
The ROLLUP Operator
The ROLLUP operator delivers aggregates and superaggregates for expressions within a GROUP BY statement. The ROLLUP operator can be used by report writers to extract statistics and summary information from results sets.(从结果集中得到精确的计算和信息概要) The cumulative aggregates can be used in reports, charts, and graphs.
The ROLLUP operator creates groupings by moving in one direction(从一个方向移动), from right to left, along the list of columns specified in the GROUP BY clause. It then applies the aggregate function to these groupings.
/././././.
Note: To produce subtotals in n dimensions (that is, n columns in the GROUP BY clause) without a ROLLUP operator, n+1 SELECT statements must be linked with UNION ALL. (如果没有rollup操作,而产生N元小计值,要使用N+1个select操作用NUION ALL连接)This makes the query execution inefficient, because each of the SELECT statements causes table access. The ROLLUP operator gathers its results with just one table access. The ROLLUP operator is useful if there are many columns involved(许多与列有关的) in producing the subtotals.
--ROLLUP 操作的例子
SELECT department_id, job_id, SUM(salary)
FROM employees
WHERE department_id < 60
GROUP BY ROLLUP(department_id, job_id);
DEPARTMENT_ID JOB_ID SUM(SALARY)
------------- ---------- -----------
10 AD_ASST 4400
10 4400
20 MK_MAN 13000
20 MK_REP 6000
20 19000
30 PU_MAN 11000
30 PU_CLERK 13900
30 24900
40 HR_REP 6500
40 6500
50 ST_MAN 36400
50 SH_CLERK 64300
50 ST_CLERK 55700
50 156400
211200
Example of a ROLLUP Operator
In the example in the slide:
:: Total salaries for every job ID within a department for those departments whose department ID is less than 60 are displayed by the GROUP BY clause (labeled 1)
1.选出条件department_id<60的,并且按GROUP BY子句显示.
10 AD_ASST 4400
20 MK_MAN 13000
20 MK_REP 6000
30 PU_MAN 11000
30 PU_CLERK 13900
40 HR_REP 6500
50 ST_MAN 36400
50 SH_CLERK 64300
50 ST_CLERK 55700
:: The ROLLUP operator displays:
-Total salary for those departments whose department ID is less than 60 (labeled 2)显示每组部门ID<60的总的salary.
10 4400
20 19000
30 24900
40 6500
50 156400
-Total salary for all departments whose department ID is less than 60, irrespective of the job IDs (labeled 3)(不考虑JOB_ID 号的总的salary).
211200
:: All rows indicated as 1 are regular rows and all rows indicated as 2 and 3 are superaggregate rows.所有行将被显示:第一步是一般的行,在第二,三步将显示聚集行.
The ROLLUP operator creates subtotals that roll up from the most detailed level to a grand total, following the grouping list specified in the GROUP BY clause. First it calculates the standard aggregate values for the groups specified in the GROUP BY clause (in the example, the sum of salaries grouped on each job within a department). Then it creates progressively higher-level subtotals, moving from right to left through the list of grouping columns. (In the preceding example, the sum of salaries for each department is calculated, followed by the sum of salaries for all departments.)
/./././././
:: Given n expressions in the ROLLUP operator of the GROUP BY clause, the operation results in n + 1 = 2 + 1 = 3 groupings.
:: Rows based on the values of the first n expressions are called rows or regular rows and the others are called superaggregate rows.
--CUBE操作
SELECT [column,] group_function(column)...
FROM table
[WHERE condition]
[GROUP BY [CUBE] group_by_expression]
[HAVING having_expression]
[ORDER BY column];
:: CUBE 是 GROUP BY 子句的扩展
:: 能够用 CUBE 操作产生带单个 SELECT 语句的交叉表值
The CUBE Operator
The CUBE operator is an additional switch in the GROUP BY clause in a SELECT statement. The CUBE operator can be applied to all aggregate functions(能够用于所有聚集函数), including AVG, SUM, MAX, MIN, and COUNT. It is used to produce results sets that are typically used for cross-tabular reports. While ROLLUP produces only a fraction of possible subtotal combinations, CUBE produces subtotals for all possible combinations of groupings specified in the GROUP BY clause, and a grand total.
The CUBE operator is used with an aggregate function to generate additional rows in a results set. Columns included in the GROUP BY clause are cross-referenced to produce a superset of groups. The aggregate function specified in the select list is applied to these groups to produce summary values for the additional superaggregate rows. The number of extra groups in the results set is determined by the number of columns included in the GROUP BY clause.
In fact, every possible combination of the columns or expressions in the GROUP BY clause is used to produce superaggregates.
If you have n columns or expressions in the GROUP BY clause, there will be 2n possible superaggregate combinations. Mathematically, these combinations form an n-dimensional cube, which is how the operator got its name.
By using application or programming tools, these superaggregate values can then be fed into charts and graphs that convey results and relationships visually and effectively.
--CUBE 操作: 例子
SQL> select department_id,job_id,sum(salary)
2 from employees
3 where department_id<60
4 group by cube(department_id,job_id)
5 order by department_id;
DEPARTMENT_ID JOB_ID SUM(SALARY)
------------- ---------- -----------
10 AD_ASST 4400
10 4400
20 MK_MAN 13000
20 MK_REP 6000
20 19000
30 PU_CLERK 13900
30 PU_MAN 11000
30 24900
40 HR_REP 6500
40 6500
50 SH_CLERK 64300
50 ST_CLERK 55700
50 ST_MAN 36400
50 156400
AD_ASST 4400
HR_REP 6500
MK_MAN 13000
MK_REP 6000
PU_CLERK 13900
PU_MAN 11000
SH_CLERK 64300
ST_CLERK 55700
ST_MAN 36400
211200
Example of a CUBE Operator
The output of the SELECT statement in the example can be interpreted as follows:
:: The total salary for every job within a department (for those departments whose department ID is less than 60) is displayed by the GROUP BY clause (labeled 1)
10 AD_ASST 4400
20 MK_MAN 13000
20 MK_REP 6000
30 PU_MAN 11000
30 PU_CLERK 13900
40 HR_REP 6500
50 ST_MAN 36400
50 SH_CLERK 64300
50 ST_CLERK 55700
:: The total salary for those departments whose department ID is less than 60 (labeled 2)
10 4400
20 19000
30 24900
40 6500
50 156400
:: The total salary for every job irrespective of the department (labeled 3)
(每个job无关的department的总的salary)
HR_REP 6500
MK_MAN 13000
MK_REP 6000
PU_MAN 11000
ST_MAN 36400
AD_ASST 4400
PU_CLERK 13900
SH_CLERK 64300
ST_CLERK 55700
:: Total salary for those departments whose department ID is less than 60, irrespective of the job titles (labeled 4)
211200
In the preceding example, all rows indicated as 1 are regular rows, all rows indicated as 2 and 4 are superaggregate rows, and all rows indicated as 3 are cross-tabulation values.
The CUBE operator has also performed the ROLLUP operation to display the subtotals for those departments whose department ID is less than 60 and the total salary for those departments whose department ID is less than 60, irrespective of the job titles. Additionally, the CUBE operator displays the total salary for every job irrespective of the department.
Note: Similar to the ROLLUP operator, producing subtotals in n dimensions (that is, n columns in the GROUP BY clause) without a CUBE operator requires 2n SELECT statements to be linked with UNION ALL. Thus, a report with three dimensions requires 2的3次方= 8 SELECT statements to be linked with UNION ALL.
--GROUPING 函数
SELECT [column,] group_function(column) . ,
GROUPING(expr)
FROM table
[WHERE condition]
[GROUP BY [ROLLUP][CUBE] group_by_expression]
[HAVING having_expression]
[ORDER BY column];
:: GROUPING 函数既能够被用于 CUBE 操作，也可以被用于 ROLLUP 操作
:: 用 GROUPING 函数模拟能够发现在一行中的构成小计的分组
:: 用 GROUPING 函数，你能够从 ROLLUP 或 CUBE 创建的空值中区分存储的 NULL 值
:: GROUPING 函数返回 0 或 1
The GROUPING Function
The GROUPING function can be used with either the CUBE or ROLLUP operator to help you understand how a summary value has been obtained.
././././.
The GROUPING function uses a single column as its argument(使用单一列). The expr(表达式) in the GROUPING function must match one of the expressions in the GROUP BY clause. The function returns a value of 0 or 1.
The values returned by the GROUPING function(被这个函数返回的值经常用于) are useful to:
:: Determine the level of aggregation of a given subtotal; that is, the group or groups on which the subtotal is based//确定组的小计值
:: Identify whether a NULL value in the expression column of a row of the result set indicates: ././././././识别在结果集的列的表达式列中NULL值的说明:
-A NULL value from the base table (stored NULL value)
-A NULL value created by ROLLUP/CUBE (as a result of a group function on that expression)
././././././
A value of 0 returned by the GROUPING function based on an expression indicates one of the following:
:: The expression has been used to calculate the aggregate value.
表达式已计算出聚集值.
:: The NULL value in the expression column is a stored NULL value.
表达式中的NULL值是存储的NULL值
A value of 1 returned by the GROUPING function based on an expression indicates one of the following:
:: The expression has not been used to calculate the aggregate value.
:: The NULL value in the expression column is created by ROLLUP or CUBE as a result of grouping.
--GROUPING 函数: 例子
SELECT department_id DEPTID, job_id JOB,
SUM(salary),
GROUPING(department_id) GRP_DEPT,
GROUPING(job_id) GRP_JOB
FROM employees
WHERE department_id < 50
GROUP BY ROLLUP(department_id, job_id);
DEPTID JOB SUM(SALARY) GRP_DEPT GRP_JOB
---------- ---------- ----------- ---------- ----------
10 AD_ASST 4400 0 0 //
10 4400 0 1 //
20 MK_MAN 13000 0 0
20 MK_REP 6000 0 0
20 19000 0 1
30 PU_MAN 11000 0 0
30 PU_CLERK 13900 0 0
30 24900 0 1
40 HR_REP 6500 0 0
40 6500 0 1
54800 1 1 //
SQL> SELECT department_id DEPTID, job_id JOB,
2 SUM(salary),
3 GROUPING(department_id) GRP_DEPT
4 FROM employees
5 WHERE department_id < 50
6 GROUP BY ROLLUP(department_id, job_id);
DEPTID JOB SUM(SALARY) GRP_DEPT
---------- ---------- ----------- ----------
10 AD_ASST 4400 0
10 4400 0
20 MK_MAN 13000 0
20 MK_REP 6000 0
20 19000 0
30 PU_MAN 11000 0
30 PU_CLERK 13900 0
30 24900 0
40 HR_REP 6500 0
40 6500 0
54800 1
Example of a GROUPING Function
In the example in the slide, consider the summary value 4400 in the first row (labeled 1). This summary value is the total salary for the job ID of AD_ASST within department 10. To calculate this summary value, both the columns DEPARTMENT_ID and JOB_ID have been taken into account. Thus a value of 0 is returned for both the expressions GROUPING(department_id) and GROUPING(job_id).
Consider the summary value 4400 in the second row (labeled 2). This value is the total salary for department 10 and has been calculated by taking into account the column DEPARTMENT_ID; thus a value of 0 has been returned by GROUPING(department_id). Because the column JOB_ID has not been taken into account to calculate this value, a value of 1 has been returned for GROUPING(job_id). You can observe similar output in the fifth row.
In the last row, consider the summary value 23400 (labeled 3). This is the total salary for those departments whose department ID is less than 50 and all job titles. To calculate this summary value, neither of the columns DEPARTMENT_ID and JOB_ID have been taken into account. Thus a value of 1 is returned for both the expressions GROUPING(department_id) and GROUPING(job_id).
//插的,老师讲解:
Instructor Note
Explain that if the same example is run with the CUBE operator, it returns a results set that has 1 for GROUPING(department_id) and 0 for GROUPING(job_id) in the cross-tabulation rows, because the subtotal values are the result of grouping on job irrespective of department number.
SQL> SELECT department_id DEPTID, job_id JOB,
2 SUM(salary),
3 GROUPING(department_id) GRP_DEPT,
4 GROUPING(job_id) GRP_JOB
5 FROM employees
6 WHERE department_id < 50
7 GROUP BY CUBE(department_id, job_id);
8 ORDER BY department_id /deptid/DEPTID/1 ;
DEPTID JOB SUM(SALARY) GRP_DEPT GRP_JOB
---------- ---------- ----------- ---------- ----------
10 AD_ASST 4400 0 0
10 4400 0 1
20 MK_MAN 13000 0 0
20 MK_REP 6000 0 0
20 19000 0 1
30 PU_CLERK 13900 0 0
30 PU_MAN 11000 0 0
30 24900 0 1
40 HR_REP 6500 0 0
40 6500 0 1
AD_ASST 4400 1 0
HR_REP 6500 1 0
MK_MAN 13000 1 0
MK_REP 6000 1 0
PU_CLERK 13900 1 0
PU_MAN 11000 1 0
54800 1 1
--分组集合
:: GROUPING SETS 是 GROUP BY 子句更进一步的扩展
:: 你能够用 GROUPING SETS 在同一查询中定义多重分组
:: Oracle 服务器计算在 GROUPING SETS 子句中指定的所有分组，并且用 UNION ALL 操作组合单个的分组结果
:: 分组集合的效率：
-对基表仅进行一个查询
-不需要写复杂的 UNION 语句
-GROUPING SETS 有更多的元素，更好的执行性能
GROUPING SETS
GROUPING SETS are a further extension of the GROUP BY clause that let you specify multiple groupings of data(可以指定多个组的数据). Doing so facilitates efficient aggregation and hence facilitates analysis of data across multiple dimensions.
A single SELECT statement can now be written using 'GROUPING SETS' to specify various groupings (that can also include ROLLUP or CUBE operators), rather than multiple SELECT statements combined by UNION ALL operators. For example, you can say:
SELECT department_id, job_id, manager_id, AVG(salary)
FROM employees
GROUP BY
GROUPING SETS
((department_id, job_id, manager_id),
(department_id, manager_id),(job_id, manager_id));
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
SA_REP 149 7000
10 AD_ASST 101 4400
20 MK_MAN 100 13000
20 MK_REP 201 6000
30 PU_MAN 100 11000
30 PU_CLERK 114 2780
40 HR_REP 101 6500
50 ST_MAN 100 7280
50 SH_CLERK 120 2900
50 ST_CLERK 120 2625
50 SH_CLERK 121 3675
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
50 ST_CLERK 121 2675
50 SH_CLERK 122 3200
50 ST_CLERK 122 2700
50 SH_CLERK 123 3475
50 ST_CLERK 123 3000
50 SH_CLERK 124 2825
50 ST_CLERK 124 2925
60 IT_PROG 102 9000
60 IT_PROG 103 4950
70 PR_REP 101 10000
80 SA_MAN 100 12200
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
80 SA_REP 145 8500
80 SA_REP 146 8500
80 SA_REP 147 7766.66667
80 SA_REP 148 8650
80 SA_REP 149 8600
90 AD_PRES 24000
90 AD_VP 100 17000
100 FI_MGR 101 12000
100 FI_ACCOUNT 108 7920
110 AC_MGR 101 12000
110 AC_ACCOUNT 205 8300
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
149 7000
10 101 4400
20 100 13000
20 201 6000
30 100 11000
30 114 2780
40 101 6500
50 100 7280
50 120 2762.5
50 121 3175
50 122 2950
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
50 123 3237.5
50 124 2875
60 102 9000
60 103 4950
70 101 10000
80 100 12200
80 145 8500
80 146 8500
80 147 7766.66667
80 148 8650
80 149 8600
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
90 24000
90 100 17000
100 101 12000
100 108 7920
110 101 12000
110 205 8300
AD_VP 100 17000
AC_MGR 101 12000
FI_MGR 101 12000
HR_REP 101 6500
MK_MAN 100 13000
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
MK_REP 201 6000
PR_REP 101 10000
PU_MAN 100 11000
SA_MAN 100 12200
SA_REP 145 8500
SA_REP 146 8500
SA_REP 147 7766.66667
SA_REP 148 8650
SA_REP 149 8333.33333
ST_MAN 100 7280
AD_ASST 101 4400
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
AD_PRES 24000
IT_PROG 102 9000
IT_PROG 103 4950
PU_CLERK 114 2780
SH_CLERK 120 2900
SH_CLERK 121 3675
SH_CLERK 122 3200
SH_CLERK 123 3475
SH_CLERK 124 2825
ST_CLERK 120 2625
ST_CLERK 121 2675
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
ST_CLERK 122 2700
ST_CLERK 123 3000
ST_CLERK 124 2925
AC_ACCOUNT 205 8300
FI_ACCOUNT 108 7920
./././././
This statement calculates aggregates over three groupings:
(department_id, job_id, manager_id), (department_id, manager_id)
and (job_id, manager_id)
Without this enhancement in Oracle9i, multiple queries combined together with UNION ALL are required to get the output of the preceding SELECT statement. A multiquery approach is inefficient, for it requires multiple scans of the same data.
GROUPING SETS (continued)
Compare the preceding statement with this alternative:
SELECT department_id, job_id, manager_id, AVG(salary)
FROM employees
GROUP BY CUBE(department_id, job_id, manager_id);
The preceding statement computes all the 8 (2 *2 *2) groupings, though only the groups (department_id, job_id, manager_id), (department_id, manager_id) and (job_id, manager_id)are of interest to you.
SELECT department_id, job_id, manager_id, AVG(salary)
FROM employees
GROUP BY //移过来.
GROUPING SETS
((department_id, job_id, manager_id),
(department_id, manager_id),(job_id, manager_id));
Another alternative is the following statement: //结果与上面的一样
SELECT department_id, job_id, manager_id, AVG(salary)
FROM employees
GROUP BY department_id, job_id, manager_id
UNION ALL
SELECT department_id, NULL, manager_id, AVG(salary)
FROM employees
GROUP BY department_id, manager_id
UNION ALL
SELECT NULL, job_id, manager_id, AVG(salary)
FROM employees
GROUP BY job_id, manager_id;
This statement requires three scans of the base table, making it inefficient.
CUBE and ROLLUP can be thought of as grouping sets with very specific semantics. The following equivalencies show this fact:
CUBE (a,b,c) GROUPING SETS //2的3次方
is equivalent to ((a,b,c),(a,b),(a,c),(b,c),(a),(b),(c)())
ROLLUP(a,b,c) GROUPING SETS
is equivalent to ((a,b,c),(a,b),(a),())
--GROUPING SETS:例子
SELECT department_id, job_id,
manager_id,avg(salary)
FROM employees
GROUP BY GROUPING SETS
((department_id,job_id), (job_id,manager_id));
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
SA_REP 7000
10 AD_ASST 4400
20 MK_MAN 13000
20 MK_REP 6000
30 PU_MAN 11000
30 PU_CLERK 2780
40 HR_REP 6500
50 ST_MAN 7280
50 SH_CLERK 3215
50 ST_CLERK 2785
60 IT_PROG 5760
70 PR_REP 10000
80 SA_MAN 12200
80 SA_REP 8396.55172
90 AD_VP 17000
90 AD_PRES 24000
100 FI_MGR 12000
100 FI_ACCOUNT 7920
110 AC_MGR 12000
110 AC_ACCOUNT 8300
AD_VP 100 17000
AC_MGR 101 12000
FI_MGR 101 12000
HR_REP 101 6500
MK_MAN 100 13000
MK_REP 201 6000
PR_REP 101 10000
PU_MAN 100 11000
SA_MAN 100 12200
SA_REP 145 8500
SA_REP 146 8500
SA_REP 147 7766.66667
SA_REP 148 8650
...
GROUPING SETS: Example
The query in the slide calculates aggregates over two groupings. The table is divided into the following groups:
-Department ID, Job ID
-Job ID, Manager ID
//每个组的平均salary要被计算.
The average salaries for each of these groups are calculated. The results set displays average salary for each of the two groups.
In the output, the group marked as 1 can be interpreted as:(标记为1的说明)
:: The average salary of all employees with the job ID AD_ASST in the department 10 is 4400.
:: The average salary of all employees with the job ID MK_MAN in the department 20 is 13000.
:: The average salary of all employees with the job ID MK_REP in the department 20 is 6000.
:: The average salary of all employees with the job ID ST_CLERK in the department 50 is 2925 and so on.
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
10 AD_ASST 4400
20 MK_MAN 13000 1
20 MK_REP 6000
50 ST_CLERK 2925
...
...
GROUPING SETS: Example (continued)
The group marked as 2 in the output is interpreted as:
:: The average salary of all employees with the job ID FI_MGR, who report to the manager with the manager ID 101, is 12000.
:: The average salary of all employees with the job ID HR_REP, who report to the manager with the manager ID 101, is 6500, and so on.
DEPARTMENT_ID JOB_ID MANAGER_ID AVG(SALARY)
------------- ---------- ---------- -----------
FI_MGR 101 12000 2
HR_REP 101 6500
The example in the slide can also be written as:
SELECT department_id, job_id, NULL as manager_id,
AVG(salary) as AVGSAL
FROM employees
GROUP BY department_id, job_id
UNION ALL
SELECT NULL, job_id, manager_id, avg(salary) as AVGSAL
FROM employees
GROUP BY job_id, manager_id;
In the absence of an optimizer that looks across query blocks to generate the execution plan, the preceding query would need two scans of the base table, EMPLOYEES. This could be very inefficient. Hence the usage of the GROUPING SETS statement is recommended.
复合列
复合列是一个作为整体被处理的列集合
ROLLUP (a, (b,c), d)
:: 为了指定复合列，用 GROUP BY 子句来分组在圆括号内的列，因此， Oracle 服务器在进行 ROLLUP 或 CUBE 操作时将它们作为一个整体来处理
::当使用 ROLLUP 或 CUBE 时，复合列将跳过在确定级别上的集合 ././././.
Composite Columns
A composite column is a collection of columns that are treated as a unit during the computation of groupings. You specify the columns in parentheses(括号) as in the following statement:
ROLLUP (a, (b, c), d) (b,c)就是一个复合列,并且做为一个单元处理.
/././././././././././.
Here, (b,c) form a composite column and are treated as a unit. In general, composite columns are useful in ROLLUP, CUBE, and GROUPING SETS. For example, in CUBE or ROLLUP, composite columns would mean skipping aggregation across certain levels.
That is, GROUP BY ROLLUP(a, (b, c))
is equivalent to ////
GROUP BY a, b, c UNION ALL
GROUP BY a UNION ALL
GROUP BY ()
Here, (b, c) are treated as a unit and rollup will not be applied across (b, c). It is as if you have an alias, for example z, for (b, c), and the GROUP BY expression reduces to
GROUP BY ROLLUP(a, z).
Note: GROUP BY( ) is typically a SELECT statement with NULL values for the columns a and b and only the aggregate function. This is generally used for generating the grand totals.
SELECT NULL, NULL, aggregate_col
FROM <table_name>
GROUP BY ( );
(continued)
Compare this with the normal ROLLUP as in:
GROUP BY ROLLUP(a, b, c)
which would be
GROUP BY a, b, c UNION ALL ////
GROUP BY a, b UNION ALL
GROUP BY a UNION ALL
GROUP BY ().
Similarly,
GROUP BY CUBE((a, b), c)
would be equivalent to ////
GROUP BY a, b, c UNION ALL
GROUP BY a, b UNION ALL
GROUP BY c UNION ALL
GROUP By ()
The following table shows grouping sets specification and equivalent GROUP BY specification.
GROUPING SETS Satements Equivalent GROUP BY Statements
GROUP BY a UNION ALL
GROUP BY GROUPING SETS(a,b,c) GROUP BY b UNION ALL
GROUP BY c
GROUP BY GROUPING SETS(a,b,(b,c)) GROUP BY a UNION ALL
(The GROUPING SETS expression has a GROUP BY b UNION ALL
composite column) GROUP BY b,c
GROUP BY GROUPING SETS((a,b,c)) GROUP BY a,b,c
GROUP BY a UNION ALL
GROUP BY GROUPING SETS(a,(b),()) GROUP BY b UNION ALL
GROPY BY ()
GROUP BY GROUPING SETS(a,ROLLUP(b,c)) GROUP BY a UNION ALL
(The GROUPING SETS expression has a GROUP BY ROLLUP(b,c)
composite column)
复合列:例子
SELECT department_id, job_id, manager_id, SUM(salary)
FROM employees
GROUP BY ROLLUP( department_id,(job_id, manager_id));
把(job_id, manager_id)看成一个单元处理.
grouping by :
1. (department_id,job_id,manager_id)
2. (department_id)
3. ( )
Composite Columns: Example
Consider the example:
SELECT department_id, job_id,manager_id, SUM(salary)
FROM employees
GROUP BY ROLLUP( department_id,job_id, manager_id);
The preceding query results in the Oracle Server computing the following groupings:
1. (department_id, job_id, manager_id)
2. (department_id, job_id)
3. (department_id)
4. ( ) //所有的sum(salary)
DEPARTMENT_ID JOB_ID MANAGER_ID SUM(SALARY)
------------- ---------- ---------- -----------
10 AD_ASST 101 4400
10 AD_ASST 4400
10 4400
20 MK_MAN 100 13000
20 MK_MAN 13000
20 MK_REP 201 6000
20 MK_REP 6000
20 19000
30 PU_MAN 100 11000
30 PU_MAN 11000
30 PU_CLERK 114 13900
30 PU_CLERK 13900
30 24900
50 ST_MAN 100 36400
50 ST_MAN 36400
50 SH_CLERK 120 11600
50 SH_CLERK 121 14700
50 SH_CLERK 122 12800
50 SH_CLERK 123 13900
50 SH_CLERK 124 11300
50 SH_CLERK 64300
50 ST_CLERK 120 10500
50 ST_CLERK 121 10700
50 ST_CLERK 122 10800
50 ST_CLERK 123 12000
50 ST_CLERK 124 11700
50 ST_CLERK 55700
50 156400
...
...
691400
If you are just interested in grouping of lines (1), (3), and (4) in the preceding example, you cannot limit the calculation to those groupings without using composite columns. With composite columns, this is possible by treating JOB_ID and MANAGER_ID columns as a single unit while rolling up. Columns enclosed in parentheses are treated as a unit while computing ROLLUP and CUBE. This is illustrated in the example on the slide. By enclosing JOB_ID and MANAGER_ID columns in parenthesis, we indicate to the Oracle Server to treat JOB_ID and MANAGER_ID as a single unit, as a composite column.
(continued)
The example in the slide computes the following groupings:
: (department_id, job_id, manager_id)
: (department_id)
:( )
The example in the slide displays the following:
: Total salary for every department (labeled 1)
: Total salary for every department, job ID, and manager (labeled 2)
: Grand total (labeled 3)
The example in the slide can also be written as:
SELECT department_id, job_id, manager_id, SUM(salary)
FROM employees
GROUP BY department_id,job_id, manager_id
UNION ALL
SELECT department_id, TO_CHAR(NULL),TO_NUMBER(NULL), SUM(salary)
FROM employees
GROUP BY department_id
UNION ALL
SELECT TO_NUMBER(NULL), TO_CHAR(NULL),TO_NUMBER(NULL), SUM(salary)
FROM employees
GROUP BY ();
In the absence of an optimizer that looks across query blocks to generate the execution plan, the preceding query would need three scans of the base table, EMPLOYEES. This could be very inefficient. Hence, the use of composite columns is recommended.
SQL> SELECT TO_NUMBER(NULL), TO_CHAR(NULL),TO_NUMBER(NULL), SUM(salary)
2 FROM employees
3 GROUP BY ();
TO_NUMBER(NULL) T TO_NUMBER(NULL) SUM(SALARY)
--------------- - --------------- -----------
691400
SQL> select count(*),sum(salary)
2 from employees;
COUNT(*) SUM(SALARY)
---------- -----------
107 691400
SQL> select null,null,null,sum(salary)
2 from employees;
N N N SUM(SALARY)
- - - -----------
691400
SQL> select null,null,null,sum(salary)
2 FROM employees
3 GROUP BY ();
N N N SUM(SALARY)
- - - -----------
691400
连接分组
:: 连接分组提供一种简明的方式来生成有用的分组组合
:: 为了指定连接分组集合，用逗号分开多重分组集合， ROLLUP，和 CUBE 操作，以便 Oracle 服务器将它们组合在一个单个的 GROUP BY 子句中
:: 分组是每个分组集合交叉乘积的结果
GROUP BY GROUPING SETS(a, b), GROUPING SETS(c, d)
Concatenated Columns
Concatenated groupings offer a concise way to generate useful combinations of groupings. The concatenated groupings are specified simply by listing multiple grouping sets, cubes, and rollups, and separating them with commas. Here is an example of concatenated grouping sets:
GROUP BY GROUPING SETS(a, b), GROUPING SETS(c, d)
The preceding SQL defines the following groupings:
(a, c), (a, d), (b, c), (b, d)
Concatenation of grouping sets is very helpful for these reasons:
:: Ease of query development: you need not manually enumerate all groupings
//不用手动列举所有的组.
:: Use by applications: SQL generated by OLAP applications often involves concatenation of grouping sets, with each grouping set defining groupings needed for a dimension
--连接分组例子
SELECT department_id, job_id, manager_id,
SUM(salary)
FROM employees
GROUP BY department_id,ROLLUP(job_id),CUBE(manager_id);
DEPARTMENT_ID JOB_ID MANAGER_ID SUM(SALARY)
------------- ---------- ---------- -----------
SA_REP 149 7000
10 AD_ASST 101 4400
20 MK_MAN 100 13000
20 MK_REP 201 6000
30 PU_MAN 100 11000
...
80 SA_MAN 100 61000
100 FI_MGR 101 12000
100 FI_ACCOUNT 108 39600
110 AC_MGR 101 12000
110 AC_ACCOUNT 205 8300
...
149 7000
10 101 4400
20 100 13000
20 201 6000
30 100 11000
30 114 13900
50 123 25900
50 124 23000
60 102 9000
60 103 19800
...
100 101 12000
100 108 39600
110 101 12000
110 205 8300
SA_REP 7000
7000
10 AD_ASST 4400
10 4400
20 MK_MAN 13000
20 MK_REP 6000
20 19000
...
50 ST_MAN 36400
50 SH_CLERK 64300
50 ST_CLERK 55700
50 156400
60 IT_PROG 28800
60 28800
70 PR_REP 10000
70 10000
...
100 FI_ACCOUNT 39600
100 51600
110 AC_MGR 12000
110 AC_ACCOUNT 8300
110 20300
Concatenated Groupings Example
The example in the slide results in the following groupings:
-(department_id, manager_id, job_id )
-(department_id, manager_id)
-(department_id, job_id)
-(department_id)
The total salary for each of these groups is calculated.
The example in the slide displays the following:
:: Total salary for every department, job ID, manager
:: Total salary for every department, manager ID
:: Total salary for every department, job ID
:: Total salary for every department
For easier understanding, the details for the department 10 are highlighted in the output.
SUMMARY
在本课中, 您应该已经学会如何：
:: 用 ROLLUP 操作产生小计值
:: 用 CUBE 操作产生交叉表值
:: 用 GROUPING 函数指定由 ROLLUP 或 CUBE 创建的行值
:: 用 GROUPING SETS 语法定义在同一查询中的多重分组
:: 用 GROUP BY 子句以不同的方式组合表达式：
-组合列
-接分组集合
Summary
:: ROLLUP and CUBE are extensions of the GROUP BY clause.
:: ROLLUP is used to display subtotal and grand total values.//每组值+总的值
:: CUBE is used to display cross-tabulation values. ///交叉
:: The GROUPING function helps you determine whether a row is an aggregate produced by a CUBE or ROLLUP operator.
:: With the GROUPING SETS syntax, you can define multiple groupings in the same query. GROUP BY computes all the groupings specified and combines them with UNION ALL.
:: Within the GROUP BY clause, you can combine expressions in various ways:
-To specify composite columns, you group columns within parentheses so that the Oracle Server treats them as a unit while computing ROLLUP or CUBE operations.
-To specify concatenated grouping sets, you separate multiple grouping sets, ROLLUP, and CUBE operations with commas so that the Oracle Server combines them into a single GROUP BY clause. The result is a cross-product of groupings from each grouping set.
分享到：
发表评论
-
ORACLE函数 非著名函数之单值函数
2011-07-13 18:23 1252注：N表示数字型，C表示字符型，D表示日期型，[]表示内中参数 ... -
ORACLE 著名函数之单值函数
2011-07-13 18:22 1006注：N表示数字型，C表 ... -
Oracle9i 对 DML 和 DDL 语句的扩展
2011-07-13 18:16 21INSERT 语句回顾 用 INSERT 语句添加新行 ... -
ORACLE 高级子查询
2011-07-13 18:15 10312:: 写多列子查询 :: 在返回空值时描述并解释子查询的行为 ... -
ORACLE 运算
2011-07-13 18:10 1454集合运算 集合运算 集合运算组合两个或多个部分查询的结果 ... -
ORACLE 权限和角色
2011-07-13 18:09 1720控制用户访问 在多 ... -
ORACLE 约束&&视图
2011-07-13 18:05 1526什么是约束? :: 约束强制规则在表级 :: 如果有从属关系， ... -
ORACLE 操作和管理表
2011-07-13 18:04 1359数据库对象: :: Table：存储数据 :: View： ... -
ORACLE 操纵数据&&锁&&事务
2011-07-13 18:03 1962INSERT语句语法 使用INSERT语句添加新行到表中 ... -
ORACLE 表连接总结
2011-07-13 17:59 2182多表连接 使用一个连接从多个表中查询数据 SELECT ta ... -
ORACLE 单行函数和分组统计总结Top-N
2011-07-13 17:58 5514alter session set nls_date_lang ...
相关推荐
Oracle的单值函数(字符函数，日期时间函数，数字函数，转换函数，混合函数)和分组函数整理
Oracle按不同时间分组统计的sql 如下表table1： 日期(exportDate) 数量(amount) -------------- ----------- 14-2月 -08 20 10-3月 -08 2 14-4月 -08 6 14-6月 -08 75 24-10月-09 23 14-11月-09 45 04-8月 -10 5 ...
给出在this时区=other时区的日期和时间 SQL> select to_char(sysdate,'yyyy.mm.dd hh24:mi:ss') bj_time,to_char(new_time 2 (sysdate,'PDT','GMT'),'yyyy.mm.dd hh24:mi:ss') los_angles from dual; BJ_TIME LOS_...
给出在this时区=other时区的日期和时间 SQL> select to_char(sysdate,'yyyy.mm.dd hh24:mi:ss') bj_time,to_char(new_time 2 (sysdate,'PDT','GMT'),'yyyy.mm.dd hh24:mi:ss') los_angles from dual; BJ_TIME ...
第12章 分组函数 第13章 当一个查询依赖于另一个查询时 第14章 一些复杂的技术 第15章 更改数据：插入﹑更新﹑合并和删除 第16章 decode 和case：sql中的if-then-else 第17章 创建和管理表、视图、索引...
理解单行函数和分组函数的区别； 掌握5个常用分组函数； 掌握分组 GROUP BY语句； 掌握过滤分组后的记录 HAVING 语句； 掌握SELECT语句6个子句的执行顺序； 利用分组函数和分组语句能解决常见问题;
第12章 分组函数 第13章 当一个查询依赖于另一个查询时 第14章 一些复杂的技术 第15章 更改数据：插入﹑更新﹑合并和删除 第16章 decode 和case：sql中的if-then-else 第17章 创建和管理表、视图、索引...
4.6 函数：字符串函数、数字函数和聚集函数(不是分组) 91 4.6.1 字符串函数 91 4.6.2 数字函数 92 4.6.3 聚集函数 92 4.7 日期函数(格式化的和按时间排列的) 93 4.7.1 日期函数 93 4.7.2 特殊格式的日期数据...
Bob Bryla是Oracle 9i和10g的认证专家，他在数据库设计、数据库应用程序开发、培训和Oracle数据库管理等方面拥有20多年的工作经验,他也足Dodgeville的Land'End公司的首席Internet数据库设计师和Oracle DBA. ...
作者通过总结各自多年的软件开发和教学培训经验，与大家分享了掌握Oracle SQL所独有的丰富功能的技巧所在，内容涵盖SQL执行、联结、集合、分析函数、子句、事务处理等多个方面。读者可以学习到以下几个方面的技巧：...
3.5.3 单行日期时间函数 3.5.4 单行转换函数 3.5.5 几个函数的使用举例 3.6 小结 第4章 PUSQL语言基础 4.1 PL／SQL语言简介 4.1.1 概述 4.1.2 开发调试环境 4.1.3 块的类型 4.2 程序...
第12章 分组函数 第13章 当一个查询依赖于另一个查询时 第14章 一些复杂的技术 第15章 更改数据：插入﹑更新﹑合并和删除 第16章 decode 和case：sql中的if-then-else 第17章 创建和管理表、视图、索引...
第12章 分组函数 第13章 当一个查询依赖于另一个查询时 第14章 一些复杂的技术 第15章 更改数据：插入﹑更新﹑合并和删除 第16章 decode 和case：sql中的if-then-else 第17章 创建和管理表、视图、索引...
4.6 函数：字符串函数、数字函数和聚集函数（不是分组） 4.6.1 字符串函数 4.6.2 数字函数 4.6.3 聚集函数 4.7 日期函数（格式化的和按时间排列的） 4.7.1 日期函数 4.7.2 特殊格式的日期数据类型 4.7.3 嵌套函数 ...
查看函数和过程的源代码 SQL>select text from all_source where owner=user and name=upper('&plsql_name'); 三、查看数据库的SQL 1、查看表空间的名称及大小 select t.tablespace_name, round(sum(bytes/...
SQL函数 5．1 数字函数 5．2 字符函数 5．3 日期时间函数 5．4 转换函数 5．5 集合函数 5．6 其他单行函数 5．7 分组函数 5．8 对象函数 5．9 习题第 6章 访问Oracle 6．1 检索...
日期类型 date 7字节 用于存储表中的日期和时间数据，取值范围是公元前4712年1月1日至公元9999年12月31日，7个字节分别表示世纪、年、月、日、时、分和秒 二进制数据类型 row 1~2000字节 可变长二进制数据，在具体...