SQL基础-->层次化查询(START BY ... CONNECT BY PRIOR)[转]

--======================================================

--SQL基础-->层次化查询(START BY ... CONNECT BY PRIOR)

--======================================================

层次化查询,即树型结构查询,是SQL中经常用到的功能之一,通常由根节点,父节点,子节点,叶节点组成,其语法如下:

SELECT [LEVEL] ,column,expression,...

FROM table_name

[WHERE where_clause]

[[START WITH start_condition] [CONNECT BY PRIOR prior_condition]];

LEVEL:为伪列,用于表示树的层次

start_condition：层次化查询的起始条件

prior_condition：定义父节点和子节点之间的关系

--使用start with ...connect by prior 从根节点开始遍历

SQL> select empno,mgr,ename,job from emp

2  start with empno = 7839

3  connect by prior empno = mgr;

EMPNO        MGR ENAME      JOB

---------- ---------- ---------- ---------

7839            KING       PRESIDENT

7566       7839 JONES      MANAGER

7788       7566 SCOTT      ANALYST

7876       7788 ADAMS      CLERK

7902       7566 FORD       ANALYST

7369       7902 SMITH      CLERK

7698       7839 BLAKE      MANAGER

7499       7698 ALLEN      SALESMAN

7521       7698 WARD       SALESMAN

7654       7698 MARTIN     SALESMAN

7844       7698 TURNER     SALESMAN

EMPNO        MGR ENAME      JOB

---------- ---------- ---------- ---------

7900       7698 JAMES      CLERK

7782       7839 CLARK      MANAGER

7934       7782 MILLER     CLERK

14 rows selected.

树型结构遍历过程(通过上面的查询来描述)

1).从根节点开始(即where_clause中的条件,如果为非根节点则分根节点作为根节点开始遍历,如上例empno = 7839)

2).遍历根节点(得到empno = 7839记录的相关信息)

3).判断该节点是否存在由子节点，如果则访问最左侧未被访问的子节点,转到),否则下一步

如上例中prior_condition为empno = mgr,即子节点的mgr等于父节点的empno,在此时mgr为7839的记录

4).当节点为叶节点，则访问完毕,否则,转到)

5).返回到该节点的父节点,转到)

--伪列level的使用

--注意connect by prior empno = mgr 的理解

--prior表示前一条记录,即下一条返回记录的mgr应当等于前一条记录的empno

SQL> select level,empno,mgr,ename,job from emp

2  start with ename = 'KING'

3  connect by prior empno = mgr

4  order by level;

LEVEL      EMPNO        MGR ENAME      JOB

---------- ---------- ---------- ---------- ---------

1       7839            KING       PRESIDENT

2       7566       7839 JONES      MANAGER

2       7698       7839 BLAKE      MANAGER

2       7782       7839 CLARK      MANAGER

3       7902       7566 FORD       ANALYST

3       7521       7698 WARD       SALESMAN

3       7900       7698 JAMES      CLERK

3       7934       7782 MILLER     CLERK

3       7499       7698 ALLEN      SALESMAN

3       7788       7566 SCOTT      ANALYST

3       7654       7698 MARTIN     SALESMAN

LEVEL      EMPNO        MGR ENAME      JOB

---------- ---------- ---------- ---------- ---------

3       7844       7698 TURNER     SALESMAN

4       7876       7788 ADAMS      CLERK

4       7369       7902 SMITH      CLERK

--获得层次数

SQL> select count(distinct level) "Level" from emp

2  start with ename = 'KING'

3  connect by prior empno = mgr;

Level

----------

4

--格式化层次查询结果(使用左填充* level - 1个空格)

SQL> col Ename for a30

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename",

3    job

4  from emp

5  start with ename = 'KING'

6  connect by prior empno = mgr;

LEVEL Ename                          JOB

---------- ------------------------------ ---------

1  KING                          PRESIDENT

2    JONES                       MANAGER

3      SCOTT                     ANALYST

4        ADAMS                   CLERK

3      FORD                      ANALYST

4        SMITH                   CLERK

2    BLAKE                       MANAGER

3      ALLEN                     SALESMAN

3      WARD                      SALESMAN

3      MARTIN                    SALESMAN

3      TURNER                    SALESMAN

LEVEL Ename                          JOB

---------- ------------------------------ ---------

3      JAMES                     CLERK

2    CLARK                       MANAGER

3      MILLER                    CLERK

14 rows selected.

--从非根节点开始遍历(只需修改start with 中的条件即可)

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename",

3    job

4  from emp

5  start with ename = 'SCOTT'

6  connect by prior empno = mgr;

LEVEL Ename                          JOB

---------- ------------------------------ ---------

1  SCOTT                         ANALYST

2    ADAMS                       CLERK

--从下向上遍历(交换connect by prior中的条件即可,使用mgr = empno)

--注意connect by prior mgr = empno 的理解
    --prior表示前一条记录,即下一条返回记录的empno应当等于前一条记录的mgr

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename",

3    job

4  from emp

5  start with ename = 'SCOTT'

6  connect by prior mgr = empno;

LEVEL Ename                          JOB

---------- ------------------------------ ---------

1  SCOTT                         ANALYST

2    JONES                       MANAGER

3      KING                      PRESIDENT

--从下向上遍历(也可以将prior置于等号右边,得到相同的结果)

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename",

3    job

4  from emp

5  start with ename = 'SCOTT'

6  connect by empno = prior mgr;

LEVEL Ename                          JOB

---------- ------------------------------ ---------

1  SCOTT                         ANALYST

2    JONES                       MANAGER

3      KING                      PRESIDENT

--从层次查询中删除节点和分支

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename"

3    ,job

4  from emp

5  where ename != 'SCOTT'    --通过where子句来过滤SCOTT用户，但SCOTT的下属ADAMS并没有过滤掉

6  start with empno = 7839

7  connect by prior empno = mgr;

LEVEL Ename                JOB

---------- -------------------- ---------

1  KING                PRESIDENT

2    JONES             MANAGER

4        ADAMS         CLERK

3      FORD            ANALYST

4        SMITH         CLERK

2    BLAKE             MANAGER

3      ALLEN           SALESMAN

3      WARD            SALESMAN

3      MARTIN          SALESMAN

3      TURNER          SALESMAN

3      JAMES           CLERK

LEVEL Ename                JOB

---------- -------------------- ---------

2    CLARK             MANAGER

3      MILLER          CLERK

13 rows selected.

--通过将过滤条件由where 子句的内容移动到connect by prior 子句中过滤掉SCOTT及其下属

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename"

3    ,job

4  from emp

5  start with empno = 7839

6  connect by prior empno = mgr and ename != 'SCOTT';

LEVEL Ename                JOB

---------- -------------------- ---------

1  KING                PRESIDENT

2    JONES             MANAGER

3      FORD            ANALYST

4        SMITH         CLERK

2    BLAKE             MANAGER

3      ALLEN           SALESMAN

3      WARD            SALESMAN

3      MARTIN          SALESMAN

3      TURNER          SALESMAN

3      JAMES           CLERK

2    CLARK             MANAGER

LEVEL Ename                JOB

---------- -------------------- ---------

3      MILLER          CLERK

12 rows selected.

--在层次化查询中增加过滤条件或使用子查询

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename"

3    ,job

4  from emp

5  where sal > 2500

6  start with empno = 7839

7  connect by prior empno = mgr

8  ;

LEVEL Ename                JOB

---------- -------------------- ---------

1  KING                PRESIDENT

2    JONES             MANAGER

3      SCOTT           ANALYST

3      FORD            ANALYST

2    BLAKE             MANAGER

SQL> select level,

2    lpad(' ',2 * level - 1) || ename as "Ename"

3    ,job

4  from emp

5  where sal > (select avg(sal) from emp)

6  start with empno = 7839

7  connect by prior empno = mgr ;

LEVEL Ename                JOB

---------- -------------------- ---------

1  KING                PRESIDENT

2    JONES             MANAGER

3      SCOTT           ANALYST

3      FORD            ANALYST

2    BLAKE             MANAGER

2    CLARK             MANAGER

6 rows selected.

更多参考：

Oracle 数据库实例启动关闭过程

Oracle 10g SGA 的自动化管理

使用OEM,SQL*Plus,iSQL*Plus 管理Oracle实例

Oracle实例和Oracle数据库(Oracle体系结构)

SQL 基础-->常用函数

SQL基础-->过滤和排序

SQL 基础-->SELECT 查询