1、之前kettle cdc mysql的时候使用的方案是canal+kafka+kettle，今天我们一起学习下使用kettle的插件Debezium直接cdc mysql。
注：CDC (Change Data Capture) 是一种技术，用于捕获和同步数据库中的更改。
1）Debezium步骤解析mysql binlog日志。
2）json input步骤解析json字符串。
3）switch-case 根据op字段进行路由。
4）create、delete、update、ddl是写日志步骤，模拟后续操作。
2、Debezium步骤配置如下（只输出dml），更多属性配置参考Debezium官网。如下图所示：
3、switch-case配置如下，不做过多介绍。
4、binlog解析之后的dml语句对应的json字符串结构
CREATE
{"ts_ms": 1718273338000,"db": "test","table": "t1","op": "CREATE","after": "{\"id\":6,\"name\":\"金刚葫芦娃\",\"createtime\":1718297698000}","pk": "{\"id\":6}"
}
UPDATE
{"ts_ms": 1718273345000,"db": "test","table": "t1","op": "UPDATE","before": "{\"id\":6,\"name\":\"金刚葫芦娃\",\"createtime\":1718297698000}","after": "{\"id\":6,\"name\":\"金刚葫芦娃plus\",\"createtime\":1718297698000}","pk": "{\"id\":6}"
}
DELETE
{"ts_ms": 1718273369000,"db": "test","table": "t1","op": "DELETE","before": "{\"id\":6,\"name\":\"金刚葫芦娃plus\",\"createtime\":1718297698000}","pk": "{\"id\":6}"
}
5、Debezium步骤配置中设置ddl为true（输出dml和ddl），更多属性配置参考Debezium官网。如下图所示：
6、binlog解析之后的ddl语句对应的json字符串结构
DROP TABLE
{"ts_ms": 1718274150072,"db": "test","table": "t3","op": "ddl","sql": "DROP TABLE `test`.`t3`"
}
CREATE TABLE
{"ts_ms": 1718274165789,"db": "test","table": "t3","op": "ddl","ddl_type": "CREATE","sql": "CREATE TABLE `t3` (\n `id` bigint NOT NULL AUTO_INCREMENT,\n `name` varchar(10) CHARACTER SET utf8mb4 COLLATE utf8mb4_general_ci DEFAULT NULL,\n `createtime` datetime DEFAULT NULL,\n PRIMARY KEY (`id`)\n ) ENGINE=InnoDB AUTO_INCREMENT=7 DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_general_ci"
}
