mysql通过@变量实现递归详细实例_数据库_开发者

1. 测试表结构

create table demo
(
    id varchar(100) not null
        primary key,
    parentId    varchar(100) not null
);
-- 测试数据
INSERT INTO demo (ID, parentID) VALUES ('1', '0');
INSERT INTO demo (ID, parentID) VALUES ('10', '9');
INSERT INTO demo (ID, parentID) VALUES ('11', '9');
INSERT INTO demo (ID, parentID) VALUES ('2', '1');
INSERT INTO demo (ID, parentID) VALUES ('3', '2');
INSERT INTO demo (ID, parentID) VALUES ('4', '3');
INSERT INTO demo (ID, parentID) VALUES ('5', '4');
INSERT INTO demo (ID, parentID) VALUES ('6', '5');
INSERT INTO demo (ID, parentID) VALUES ('7', '6');
INSERT INTO demo (ID, parentID) VALUES ('8', '1');
INSERT INTO demo (ID, parentID) VALUES ('9', '1');

2. 向下递归（包括自己）

-- 向下递归 par即为递归查询出来的所有id
SELECT
    @par                                                                          AS par,
    (SELECT @par := GROUP_CONCAT(id) FROM demo WHERE FIND_IN_SET(parentid, @par)) AS son
FROM demo_userm, (SELECT @par := '1') T
WHERE @par IS NOT NULL;
查询结果：
| par       | son     |
| 1         | 2,8,9   |
| 2,8,9     | 10,11,3 |
| 10 ,11,3  | 4       |
| 4         | 5       |
| 5         | 6       |
| 6         | 7       |
| 7         | NULL    |

3.向下递归（不包括自己）

-- 向下递归 par即为递归查询出来的所有id
SELECT
    @par                           oGZSY                                               AS par,
    (SELECT @par := GROUP_CONCAT(id) FROM demo WHERE FIND_IN_SET(parentid, @par)) AS son
FROM dejavascriptmo_userm, (SELECT @par := '1') T
WHERE @par IS NOT NULL and @par != '1';
查询结果：
| par       | son     |
| 2,8,9     | 10,11,3 |
| 10 ,11,3  | 4       |
| 4         | 5       |
| 5         | 6       |
| 6         | 7       |
| 7         | NULL    |

4. 向下递归可能存在问题，修正优化：

-- 如果存在 par = 51016 并且 son = 51016 ， 那么会出现问题，多出许多数据
SELECT
    @par                                                                          AS par,
    (SELECT @par := GROUP_CONCAT(distinct ORG_REFNO) FROM gis_udp_bank WHERE FIND_IN_SET(PNBRN_ORG_REFNO, @par)) AS son
FROM gis_udp_bank, (SELECT @par := '51016') T
WHERE @par IS NOT NULL;
-- 优化后：后面家条件： !FIND_IN_SET(ORG_REFNO, @par)
SELECT
    @par                                                                          AS par,
    (SELECT @par := GROUP_CONCAT(distinct ORG_REFNO) FROM gis_udp_bank WHERE FIND_IN_SET(PNBRN_ORG_REFNO, @par)
        and !FIND_IN_SET(ORG_REFNO, @par)
        ) AS son
FROM gis_udp_bank, oGZSY(SELECT @par := '51016') T
WHERE @par IS NOT NULL;

5. 向上递归

SELECT
    @par                                                                     AS par,
    (SELECT @par := GROUP_CONCAT(NAME) FROM demo_userm WHERE LOGONID = @par) AS son
FROM demo_userm, (SELECT @par := '10') T
WHERE @par IS NOT NULL;
par即为递归查询结果
查询结果：
| par | son |
| 10  | 9   |
| 9   | 1   |
| 1   | 0   |
| 0   | NULL|

所用函数介绍：

1、GROUP_CONCAT([distinct] colName)

一般用于取ID集合。
select  * from demo_userm t where NAME = 1;
| LOGONID | NAME |
| 2       | 1    |
| 8       | 1    |
| 9       | 1    |
select  GROUP_CONCAT(LOGONID) from demo_userm t where NAME = 1;
| GROUP\_CONCAT\(LOGONID\) |
| 2,8,9 |
-- 不去重取name集合
select  GROUP_CONCAT(NAME) from demo_userm t where NAME in (1,2,3,4,5,6,7);
| GROUP\_CONCAT\(NAME\) |
| 1,2,3,4,5,6,1,1 |
-- 去重再取name集合
select  GROUP_CONCAT(distinct NAME) from demo_userm t where NAME in (1,2,3,4,5,6,7);
| GROUP\_CONCAT\(distinct NAME\) |
| 1,2,3,4,5,6 |

2、FIND_IN_SET（colName, strs）

select * from demo_userm where  FIND_IN_SET(LOGONID, ('1,2,3,4'));
| LOGONID | NAME |
| 1       | 0    |
| 2       | 1    |
| 3       | 2    |
| 4       | 3    |
select * from demo_userm where  
FIND_Ihttp://www.devze.comN_SET(LOGONID, (select  GROUP_CONCAT(LOGONID) from demo_userm t where NAME = 1));
| LOGONID | NAME |
| 2       | 1    |
| 8       | 1    |
| 9       | 1    |