MySQL 列出相关依赖

DROP TABLE IF EXISTS test.job_depend;
# 创建测试表
CREATE TABLE `job_depend` (
  `sn_id` int(11) NOT NULL AUTO_INCREMENT COMMENT '行号',
  `job_id` varchar(20) DEFAULT NULL COMMENT '作业ID',
  `depend_job_id` varchar(20) DEFAULT NULL COMMENT '依赖作业',
  PRIMARY KEY (`sn_id`)
) ENGINE=InnoDB AUTO_INCREMENT=7 DEFAULT CHARSET=utf8 COMMENT='作业依赖';

# 插入测试数据
INSERT INTO test.job_depend (job_id, depend_job_id) VALUES ('b', 'a');
INSERT INTO test.job_depend (job_id, depend_job_id) VALUES ('c', 'b');
INSERT INTO test.job_depend (job_id, depend_job_id) VALUES ('c', 's');
INSERT INTO test.job_depend (job_id, depend_job_id) VALUES ('s', 'c');

# 若存储过程存在则删除
DROP PROCEDURE IF EXISTS dep;
# 创建一个存储过程
CREATE PROCEDURE dep(jobId VARCHAR(21845))
  # 开始内容
  BEGIN
    # 定义一个变量存储合并后的字符串（逗号分隔)
    DECLARE childs VARCHAR(21845);
    # 定义一个变量存储当前查到的字符串(逗号分隔)
    DECLARE ids VARCHAR(21845);
    # 初始化字符串
    SET childs = '';

    # 逗号分隔拼接字符串，支持查到多个
    # INTO 放入 JobId
    # 拼接 % 使自带 like 模糊查找
    SELECT DISTINCT group_concat(job_id)
    INTO jobId
    FROM job_depend
    WHERE job_id LIKE concat('%', jobId, '%');

    SET ids = jobId;

    # 当前查到的字符串不为空时
    WHILE ids IS NOT NULL DO
      # 拼接字符串
      SET childs = concat(ids, ',', childs);

      # SELECT 不重复 拼接字段，默认逗号分隔
      # INTO 放入 当前查找的字符串
      # WHERE (查找值, 字符串集合) 且没有在合并字符串
      SELECT DISTINCT group_concat(depend_job_id)
      INTO ids
      FROM job_depend
      WHERE find_in_set(job_id, ids) > 0
            AND NOT find_in_set(depend_job_id, childs);
    END WHILE;

    SET ids = jobId;

    # 反过来查找依赖 ids 的
    WHILE ids IS NOT NULL DO
      SELECT DISTINCT group_concat(job_id)
      INTO ids
      FROM job_depend
      WHERE find_in_set(depend_job_id, ids) > 0
            AND NOT find_in_set(job_id, childs);
      IF ids IS NOT NULL THEN
        SET childs = concat(childs, ',', ids);
      END IF;
    END WHILE;

    SELECT
      # DISTINCT
      # 如果 job_id 是查询传入的 jobId
      # SQL 下的 IF 条件写法
      CASE find_in_set(job_id, jobId) > 0
      # 为真 则标记 =>
      WHEN TRUE THEN '=>'
      # 否则不显示任何内容
      ELSE '' END AS f,
      job_id,
      depend_job_id
    FROM job_depend
    WHERE find_in_set(job_id, childs)
    # 按下面依赖上面排序(需查找位置的子字符串, 大字符串)
    ORDER BY instr(job_id, jobId);
  END;

# 使用例子
CALL dep('c');

# 查询本程序
SELECT
  SPECIFIC_NAME,
  ROUTINE_TYPE
FROM information_schema.ROUTINES
WHERE SPECIFIC_NAME = 'dep';