flinkcdc抽取oracle数据
直接上代码
package flinkcdc;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.table.api.TableResult;
import org.apache.flink.table.api.bridge.java.StreamTableEnvironment;
public class FlinkCdcOracle {
public static void main(String[] args) throws Exception {
//创建flink环境
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
env.setParallelism(1);
env.disableOperatorChaining();
StreamTableEnvironment tenv = StreamTableEnvironment.create(env);
tenv.executeSql("create table employees_flink(\n" +
"EMPLOYEE_ID int , \n" +
"FIRST_NAME string, \n" +
"LAST_NAME string, \n" +
"EMAIL string, \n" +
"PHONE_NUMBER string, \n" +
"JOB_ID string, \n" +
"SALARY double, \n" +
"COMMISSION_PCT double, \n" +
"MANAGER_ID int, \n" +
"DEPARTMENT_ID int\n" +
")WITH (\n" +
" 'connector' = 'oracle-cdc', \n" +
" 'hostname' = 'localhost', \n" +
" 'port' = '1521', \n" +
" 'username' = 'test', \n" +
" 'password' = 'test', \n" +
" 'database-name' = 'ORCL', \n" +
" 'schema-name' = 'TEST', // 注意这里要大写\n" +
" 'table-name' = 'employees', \n" +
" 'debezium.log.mining.continuous.mine'='true',\n" +
" 'debezium.log.mining.strategy'='online_catalog', \n" +
" 'debezium.database.tablename.case.insensitive'='false',\n" +
" 'scan.startup.mode' = 'initial')");
TableResult tableResult = tenv.executeSql("select * from employees_flink");
tableResult.print();
}
}
参数详解
Debezium 的开发者将 “大小写不敏感” 统一定义为了 “需要将表名转换为小写,数据库名转为大写”
然而对于 Oracle 数据库，“大小写不敏感” 却意味着在内部元信息存储时，需要将表名转换为大写。
因而 Debezium 在读取到 “大小写不敏感” 的配置后，按照上述代码逻辑，只会因为尝试去读取小写的表名而报错。
解决办法：需使用 Oracle “大小写不敏感” 的特性，可以在 create 语句中加上 'debezium.database.tablename.case.insensitive'='false'
*/
'debezium.database.tablename.case.insensitive'='false',
/*
以FLINKCDC读取mysql数据为例。flinkcdc读取mysql数据分为两个阶段，
首先是快照阶段，并行读取快照数据，快照阶段按照表的主键和行数别切分为为多个是数据块，被分配到到不同的reader处理。
快照阶段支持块级别的checkpoint，当checkpoint失败后从上一次的checkpoint恢复
只有当块级别的数据块完全checkpoint之后才会进入到binlog 读取阶段
仅启动一个 task，即一个并行度。为了保证快照数据和 binlog 数据的顺序 。
inlog reader 把 binlog 的位置记录在 Flink 状态中，因此，在 binlog 阶段可以支持行级别的 checkpoint。
当表没有主键时，Incremental snapshot reading 会失败，这时需要修改配置，
把 scan.incremental.snapshot.enabled 设置为 false，使用旧的快照读取机制
*/
'scan.incremental.snapshot.enabled'='false'
版本二
package flinkcdc;
import org.apache.flink.streaming.api.CheckpointingMode;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.table.api.TableResult;
import org.apache.flink.table.api.bridge.java.StreamTableEnvironment;
public class FlinkCdcOracle {
public static void main(String[] args) throws Exception {
//创建flink环境
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
env.setParallelism(1);
StreamTableEnvironment tenv = StreamTableEnvironment.create(env);
//启用检查点,检查点设置每个5000毫秒
env.enableCheckpointing(5000);
//配置检查点保存路径
env.getCheckpointConfig().setCheckpointStorage("hdfs://hdfs-cluster/flink/flinkcheckpoint");
//设置检查点模式
env.getCheckpointConfig().setCheckpointingMode(CheckpointingMode.EXACTLY_ONCE);
//设置检查点之间的最小暂停时间1000毫秒
env.getCheckpointConfig().setMinPauseBetweenCheckpoints(1000);
//设置检查点超时时间
env.getCheckpointConfig().setCheckpointTimeout(60000);
tenv.executeSql("create table employees_flink(\n" +
"EMPLOYEE_ID int , \n" +
"FIRST_NAME string, \n" +
"LAST_NAME string, \n" +
"EMAIL string, \n" +
"PHONE_NUMBER string, \n" +
"HIRE_DATE string, \n" +
"JOB_ID string, \n" +
"SALARY double, \n" +
"COMMISSION_PCT double, \n" +
"MANAGER_ID int, \n" +
"DEPARTMENT_ID int\n" +
")WITH (\n" +
" 'connector' = 'oracle-cdc', \n" +
" 'hostname' = 'localhost', \n" +
" 'port' = '1521', \n" +
" 'username' = 'test', \n" +
" 'password' = 'test', \n" +
" 'database-name' = 'ORCL', \n" +
" 'schema-name' = 'TEST', // 注意这里要大写\n" +
" 'table-name' = 'employees', \n" +
" 'debezium.log.mining.continuous.mine'='true',\n" +
" 'debezium.log.mining.strategy'='online_catalog', \n" +
" 'debezium.database.tablename.case.insensitive'='false',\n" +
" 'scan.startup.mode' = 'initial')");
TableResult tableResult = tenv.executeSql("select * from employees_flink");
tableResult.print();
}
}注意此时运行会如下错误：
侦听器当前不知道连接描述符中请求的服务。说明监听器有问题，我们查看监听器配置我是本地安装的，目录是E:\app\user\product\11.2.0\dbhome_1\NETWORK\ADMIN\listener.ora
进行修改
然后在E:\app\user\product\11.2.0\dbhome_1\NETWORK\ADMIN目录下shift+powershell 执行如下命令： lsnrctl reload
然后oracle重启服务（自己私下测试，是否重启不确定），重新启动java的demo即可
报错解决连接：
上述的cdc解决大部分的数据类型问题，但是Date类型的数据没有解决。以下重点解决date数据类型。
报错如下
Caused by: java.time.DateTimeException: Invalid value for Year (valid values -999999999 - 999999999)
不知道咋解决的使用string代替date，反正没报错，运行出来了
package flinkcdc;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.table.api.TableResult;
import org.apache.flink.table.api.bridge.java.StreamTableEnvironment;
public class FlinkCdcOracle {
public static void main(String[] args) throws Exception {
//创建flink环境
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
env.setParallelism(1);
env.disableOperatorChaining();
StreamTableEnvironment tenv = StreamTableEnvironment.create(env);
tenv.executeSql("create table employees_flink(\n" +
"EMPLOYEE_ID int , \n" +
"FIRST_NAME string, \n" +
"LAST_NAME string, \n" +
"EMAIL string, \n" +
"PHONE_NUMBER string, //原本是Date类型，报错替换成String \n" +
"HIRE_DATE string, \n" +
"JOB_ID string, \n" +
"SALARY double, \n" +
"COMMISSION_PCT double, \n" +
"MANAGER_ID int, \n" +
"DEPARTMENT_ID int\n" +
")WITH (\n" +
" 'connector' = 'oracle-cdc', \n" +
" 'hostname' = 'localhost', \n" +
" 'port' = '1521', \n" +
" 'username' = 'test', \n" +
" 'password' = 'test', \n" +
" 'database-name' = 'ORCL', \n" +
" 'schema-name' = 'TEST', // 注意这里要大写\n" +
" 'table-name' = 'employees', \n" +
" 'debezium.log.mining.continuous.mine'='true',\n" +
" 'debezium.log.mining.strategy'='online_catalog', \n" +
" 'debezium.database.tablename.case.insensitive'='false',\n" +
" 'scan.startup.mode' = 'initial')");
TableResult tableResult = tenv.executeSql("select * from employees_flink");
tableResult.print();
}
}
本文章为转载内容，我们尊重原作者对文章享有的著作权。如有内容错误或侵权问题，欢迎原作者联系我们进行内容更正或删除文章。
