程序员人生网站导航

Oracle 递归查询详解

栏目：数据库应用时间：2015-02-27 07:53:43

创建表和主外键

CREATE TABLE SC_DISTRICT ( ID NUMBER(10) NOT NULL, PARENT_ID NUMBER(10), NAME VARCHAR2(255 BYTE) NOT NULL ); ALTER TABLE SC_DISTRICT ADD ( CONSTRAINT SC_DISTRICT_PK PRIMARY KEY (ID)); ALTER TABLE SC_DISTRICT ADD ( CONSTRAINT SC_DISTRICT_R01 FOREIGN KEY (PARENT_ID) REFERENCES SC_DISTRICT (ID) );

插入数据

insert into A_DISTRICT (id, parent_id, name) values (1, null, '河南省'); insert into A_DISTRICT (id, parent_id, name) values (2, null, '山东省'); insert into A_DISTRICT (id, parent_id, name) values (3, 1, '郑州市'); insert into A_DISTRICT (id, parent_id, name) values (4, 1, '开封市'); insert into A_DISTRICT (id, parent_id, name) values (5, 1, '商丘市'); insert into A_DISTRICT (id, parent_id, name) values (6, 3, '中原区'); insert into A_DISTRICT (id, parent_id, name) values (7, 3, '金水区'); insert into A_DISTRICT (id, parent_id, name) values (8, 3, '27区'); insert into A_DISTRICT (id, parent_id, name) values (9, 3, '城管回族区'); insert into A_DISTRICT (id, parent_id, name) values (10, 3, '上街区'); insert into A_DISTRICT (id, parent_id, name) values (11, 3, '惠济区'); insert into A_DISTRICT (id, parent_id, name) values (12, 4, '开封县'); insert into A_DISTRICT (id, parent_id, name) values (13, 4, '杞县'); insert into A_DISTRICT (id, parent_id, name) values (14, 4, '兰考县'); insert into A_DISTRICT (id, parent_id, name) values (15, 4, '通许县'); insert into A_DISTRICT (id, parent_id, name) values (16, 13, '城关镇'); insert into A_DISTRICT (id, parent_id, name) values (17, 13, '5里河镇'); insert into A_DISTRICT (id, parent_id, name) values (18, 13, '邢口镇'); insert into A_DISTRICT (id, parent_id, name) values (19, 13, '柿园乡'); insert into A_DISTRICT (id, parent_id, name) values (20, 13, '城郊乡');

生成数据以下

递归语法和主要使用函数

select * from 表 start with 条件入口 connect by prior id = parent_id（向下）

select * from 表 start with 条件入口 connect by prior parent_id = id（向上）

如果不写prior，表明前序遍历未指明遍历方向，所以不进行递归，习惯上会在语句后面加上order by排序和group by分组

connect_by_root(列名) 查询根节点

connect_by_isleaf 查询是不是叶子节点，0是，1不是

sys_connect_by_path 查询递归路径

实战

查询根节点

select id, parent_id, name from a_district start with parent_id is null connect by id = parent_id order by id

递归查询杞县的所有上级节点（包括当前节点）

select id, parent_id, name from a_district start with name='杞县' connect by prior parent_id = id order by id

递归查询杞县的所有上级节点（包括当前节点和根节点）

select id, parent_id, name,connect_by_root(id) city_id, connect_by_root(name) city_name from a_district start with name='杞县' connect by prior parent_id = id order by id

递归查询杞县的所有子城镇乡（包括当前节点）

select id, parent_id, name from a_district start with name='杞县' connect by prior id = parent_id order by id

递归查询开封市的所有城镇乡（包括当前节点和根节点）

select id, parent_id, name, connect_by_root(id) city_id, connect_by_root(name) city_name from a_district start with name='开封市' connect by prior id = parent_id order by id

上行递归查询城郊乡深度、是不是包括叶子节点

select id, name, parent_id, level, connect_by_isleaf isleaf from a_district start with name = '城郊乡' connect by prior parent_id = id order by id

下行递归查询河南省深度、是不是包括叶子节点

select id, name, parent_id, level, connect_by_isleaf isleaf from a_district start with name = '河南省' connect by prior id = parent_id order by id

上行递归查询杞县路径

select id, name, parent_id, substr(sys_connect_by_path(name,'->'),3) name_path from a_district start with name='杞县' connect by prior parent_id = id order by id

下行递归查询河南省路径

select id, name, parent_id, substr(sys_connect_by_path(name,'->'),3) name_path from a_district start with name='河南省' connect by prior id = parent_id order by id

综合使用

select id, name, parent_id, level, connect_by_isleaf isleaf, substr(sys_connect_by_path(name,'->'),3) name_path from a_district start with name='河南省' connect by prior id = parent_id order by id

SYS_CONNECT_BY_PATH函数

在Oracle中，SYS_CONNECT_BY_PATH函数主要作用是可以把1个父节点下的所有子节点通过某个字符进行辨别，然后连接在1个列中显示。

sys_connect_by_path(字段名, 2个字段之间的连接符号)，注意这里的连接符号不要使用逗号，oracle会报错，如果1定要用，可使用replace替换1下，REPLACE(字段名,原字符,',')。

还有，这个函数使用之前必须先建立1个树，否则无用。

------分隔线----------------------------

上一篇 Google Gson的使用方法,实现Json结构的相互转换

下一篇 关于“博客之星”评选活动的意见

------分隔线----------------------------