简介

Flink CDC：

​ 解决了传统数据库实时同步的痛点， 该技术抛弃了其他第三方组件（例如Kafka等），能够实时读取Mysql master节点全量和增量数据，能够捕获所有数据的变化，同时它完全与业务解耦，运维也及其简单。具体介绍请参考：Flink_CDC搭建及简单使用 及 flink-cdc-connectors。

Apache Doris：

它是一个现代化的MPP分析型数据库产品。仅需亚秒级响应时间即可获得查询结果，有效地支持实时数据分析。Apache Doris的分布式架构非常简洁，易于运维，并且可以支持10PB以上的超大数据集。

Apache Doris可以满足多种数据分析需求，例如固定历史报表，实时数据分析，交互式数据分析和探索式数据分析等。令您的数据分析工作更加简单高效！目前Apache Doris已成为Apache顶级项目，成为许多一线大厂实时数据分析数据仓库的不二选择。具体请参考：Apache Doris 。

技术选型

​ 本次采用Flink-CDC和flink-doris-connector技术利用FLink DataStream的方式实现Mysql全量或增量数据同步至分析性数据仓库Doris中。

版本选择：

Flink	Flink-doris-connector	Flink-CDC	Doris
1.11.x	1.11.6-2.12-xx	1.0.0/1.1.0	0.13.+
1.12.x	1.12.7-2.12-xx	1.2.0/1.3.0	0.13.+
1.13.x	1.13.5-2.12-xx	1.4.0 / 2.0.* / 2.1.* / 2.2.*	0.13.+
1.14.x	1.14.4-2.12-xx	2.2.*	0.13.+

注：Flink-doris-connector版本号解读：例如1.13.5-2.12-1.0.1 表示flink 版本 1.13.5，scala 版本 2.12，connector 版本 1.0.1。

具体实现

本次代码构建选择的版本号为：

Flink	Flink-doris-connector	Flink-CDC	Doris
1.13.6	1.0.3	2.2.0	0.15.13

导入依赖：

        <!--flink cdc -->
        <dependency>
            <groupId>com.ververica</groupId>
            <artifactId>flink-connector-mysql-cdc</artifactId>
            <version>2.2.0</version>
        </dependency>


        <!-- flink  -->
        <dependency>
            <groupId>org.apache.flink</groupId>
            <artifactId>flink-java</artifactId>
            <version>1.13.6</version>
            <scope>provided</scope>
        </dependency>

        <dependency>
            <groupId>org.apache.flink</groupId>
            <artifactId>flink-clients_2.12</artifactId>
            <version>1.13.6</version>
            <scope>provided</scope>
        </dependency>


        <dependency>
            <groupId>org.apache.flink</groupId>
            <artifactId>flink-streaming-java_2.12</artifactId>
            <version>1.13.6</version>
            <scope>provided</scope>
        </dependency>

        <dependency>
            <groupId>org.apache.flink</groupId>
            <artifactId>flink-connector-jdbc_2.12</artifactId>
            <version>1.13.6</version>
            <scope>provided</scope>
        </dependency>

        <dependency>
            <groupId>org.apache.flink</groupId>
            <artifactId>flink-table-planner-blink_2.12</artifactId>
            <version>1.13.6</version>
            <scope>provided</scope>
        </dependency>


        <!-- flink doris connector -->
        <dependency>
            <groupId>org.apache.doris</groupId>
            <artifactId>flink-doris-connector-1.13_2.12</artifactId>
            <version>1.0.3</version>
        </dependency>

代码实现：

public static void main(String[] args) throws Exception {

        StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

        // enable checkpoint
        env.enableCheckpointing(3000);


        MySqlSource<String> mySqlSource = MySqlSource.<String>builder()
                .hostname("127.0.0.1")
                .port(3306)
                .databaseList("test") // set captured database
                .tableList("test.test_cdc") // set captured table
                .username("bigdata")
                .password("123456")
                .deserializer(new JsonDebeziumDeserializationSchema()).build();

        Properties pro = new Properties();
        pro.setProperty("format", "json");
        pro.setProperty("strip_outer_array", "true");

        DataStreamSource<String> dataStreamSource = env.fromSource(mySqlSource, WatermarkStrategy.noWatermarks(), "mysql source");

        dataStreamSource.setParallelism(1).print();

        /**
         * flink cdc同步数据时的数据几种数据格式：
         * insert  :{"before":null,"after":{"id":30,"name":"wangsan","age":27,"address":"北京"},"source":{"version":"1.5.4.Final","connector":"mysql","name":"mysql_binlog_source","ts_ms":1652844463000,"snapshot":"false","db":"test","sequence":null,"table":"test_cdc","server_id":1,"gtid":null,"file":"mysql-bin.000001","pos":10525,"row":0,"thread":null,"query":null},"op":"c","ts_ms":1652844463895,"transaction":null}
         * update :{"before":{"id":30,"name":"wangsan","age":27,"address":"北京"},"after":{"id":30,"name":"wangsan","age":27,"address":"上海"},"source":{"version":"1.5.4.Final","connector":"mysql","name":"mysql_binlog_source","ts_ms":1652844511000,"snapshot":"false","db":"test","sequence":null,"table":"test_cdc","server_id":1,"gtid":null,"file":"mysql-bin.000001","pos":10812,"row":0,"thread":null,"query":null},"op":"u","ts_ms":1652844511617,"transaction":null}
         * delete :{"before":{"id":25,"name":"wanger","age":26,"address":"西安"},"after":null,"source":{"version":"1.5.4.Final","connector":"mysql","name":"mysql_binlog_source","ts_ms":1652844336000,"snapshot":"false","db":"test","sequence":null,"table":"test_cdc","server_id":1,"gtid":null,"file":"mysql-bin.000001","pos":10239,"row":0,"thread":null,"query":null},"op":"d","ts_ms":1652844336733,"transaction":null}
         *
         * doris:
         * 因为doris是不能根据非主键字段做删除操作的，所以当mysql中有删除操作时，这边边采用逻辑删除的方式，将删出字段标记为已删除
         * 后期在做数据分析时，可以将这部分数据过滤掉即可。
         *  is_delete：逻辑删除标志符 0表示正常 1表示删除
         *  updateStamp：数据的落库时间
         */

        env.fromSource(mySqlSource, WatermarkStrategy.noWatermarks(), "mysql source").map(line -> {
            JSONObject lineObj = JSONObject.parseObject(line);
            //如果是insert或者update操作，则取after中的数据，将是否删除设置为0
            JSONObject data;
            if ("d".equals(lineObj.getString("op"))) {
                //如果是delete操作，则取before中的数据，将其设置为1
                data = JSONObject.parseObject(lineObj.getString("before"));
                data.put("is_delete", 1);
                data.put("updateStamp", new Date().toLocaleString());
            } else {
                data = JSONObject.parseObject(lineObj.getString("after"));
                data.put("is_delete", 0);
                data.put("updateStamp", new Date().toLocaleString());
            }
            return data.toJSONString();
        }).addSink(DorisSink.sink(DorisExecutionOptions.builder()
                        .setBatchSize(3)
                        .setBatchIntervalMs(10L)
                        .setMaxRetries(3)
                        .setStreamLoadProp(pro)
                        .setEnableDelete(true)
                        .build(),
                DorisOptions.builder()
                        .setFenodes("127.0.0.1:8030")
                        .setTableIdentifier("test_lit.test_cdc_sink")
                        .setUsername("root")
                        .setPassword("").build()));

        env.execute("Print MySQL Snapshot + Binlog");

    }

总结：

​ 通过Flink CDC和flink doris connector 技术，我们可以很简单的将Mysql中增量或全量数据同步至Doris中，链路短，时效性高，简化了传统数据同步的方式，便于维护。