当前位置：首页 > news >正文

Flink Sink的使用

news 2025/6/30 6:54:58

经过一系列Transformation转换操作后，最后一定要调用Sink操作，才会形成一个完整的DataFlow拓扑。只有调用了Sink操作，才会产生最终的计算结果，这些数据可以写入到的文件、输出到指定的网络端口、消息中间件、外部的文件系统或者是打印到控制台.

flink在批处理中常见的sink

print 打印
writerAsText 以文本格式输出
writeAsCsv 以csv格式输出
writeUsingOutputFormat 以指定的格式输出
writeToSocket 输出到网络端口
自定义连接器（addSink）

参考官网：https://ci.apache.org/projects/flink/flink-docs-release-1.13/zh/docs/dev/datastream/overview/#data-sinks

1、print

打印是最简单的一个Sink，通常是用来做实验和测试时使用。如果想让一个DataStream输出打印的结果，直接可以在该DataStream调用print方法。另外，该方法还有一个重载的方法，可以传入一个字符，指定一个Sink的标识名称，如果有多个打印的Sink，用来区分到底是哪一个Sink的输出。

以下演示了print打印，以及自定义print打印。

package com.bigdata.day03;import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.configuration.Configuration;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;public class SinkPrintDemo {public static void main(String[] args) throws Exception {//1. env-准备环境StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);DataStreamSource<String> dataStreamSource = env.socketTextStream("localhost", 8888);// 打印，普通的打印// 6> helllo world//dataStreamSource.print();dataStreamSource.addSink(new MySink());// 接着手动实现该print 打印env.execute();}static class MySink extends RichSinkFunction<String> {@Overridepublic void invoke(String value, Context context) throws Exception {// 得到一个分区号，因为要模仿print打印效果int partitionId = getRuntimeContext().getIndexOfThisSubtask() + 1;String msg = partitionId +"> " +value;System.out.println(msg);}}}

package com.bigdata.day03;import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;public class Demo01 {static class MyPrint extends RichSinkFunction<String>{private String msg;public MyPrint(){}public MyPrint(String msg){this.msg = msg;}@Overridepublic void invoke(String value, Context context) throws Exception {int partition = getRuntimeContext().getIndexOfThisSubtask();if(msg == null){System.out.println(partition+"> "+value);}else{System.out.println(msg+">>>:"+partition+"> "+value);}}}public static void main(String[] args) throws Exception {//1. env-准备环境StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);//2. source-加载数据DataStream<String> data = env.fromElements("hello", "world", "baotianman", "laoyan");//3. transformation-数据处理转换//4. sink-数据输出//data.print();//data.print("普通打印>>>");data.addSink(new MyPrint());data.addSink(new MyPrint("模仿:"));//5. execute-执行env.execute();}
}

下面的结果是WordCount例子中调用print Sink输出在控制台的结果，细心的读者会发现，在输出的单词和次数之前，有一个数字前缀，我这里是1~4，这个数字是该Sink所在subtask的Index + 1。有的读者运行的结果数字前缀是1~8，该数字前缀其实是与任务的并行度相关的，由于该任务是以local模式运行，默认的并行度是所在机器可用的逻辑核数即线程数，我的电脑是2核4线程的，所以subtask的Index范围是0~3，将Index + 1，显示的数字前缀就是1~4了。这里在来仔细的观察一下运行的结果发现：相同的单词输出结果的数字前缀一定相同，即经过keyBy之后，相同的单词会被shuffle到同一个subtask中，并且在同一个subtask的同一个组内进行聚合。一个subtask中是可能有零到多个组的，如果是有多个组，每一个组是相互独立的，累加的结果不会相互干扰。

sum之后的：

1> hello 3

2> world 4

汇总之前，keyBy之后

1> hello 1

2、writerAsText 以文本格式输出

该方法是将数据以文本格式实时的写入到指定的目录中，本质上使用的是TextOutputFormat格式写入的。每输出一个元素，在该内容后面同时追加一个换行符，最终以字符的形式写入到文件中，目录中的文件名称是该Sink所在subtask的Index + 1。该方法还有一个重载的方法，可以额外指定一个枚举类型的参数writeMode，默认是WriteMode.NO_OVERWRITE，如果指定相同输出目录下有相同的名称文件存在，就会出现异常。如果是WriteMode.OVERWRITE，会将以前的文件覆盖。

package com.bigdata.day03;import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.core.fs.FileSystem;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;public class SinkTextDemo {public static void main(String[] args) throws Exception {//1. env-准备环境StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);env.setParallelism(2);DataStreamSource<String> dataStreamSource = env.socketTextStream("localhost", 8880);// 写入到文件的时候，OVERWRITE 模式是重写的意思，假如以前有结果直接覆盖// 如果并行度为1 ,最后输出的结果是一个文件，假如并行度 > 1 最后的结果是一个文件夹，文件夹中的文件名是 分区号（任务号）dataStreamSource.writeAsText("F:\\BD230801\\FlinkDemo\\datas\\result", FileSystem.WriteMode.OVERWRITE);env.execute();}
}

package com.bigdata.day03;import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.core.fs.FileSystem;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;public class Demo02 {public static void main(String[] args) throws Exception {//1. env-准备环境StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);//2. source-加载数据//3. transformation-数据处理转换//4. sink-数据输出//DataStreamSource<String> streamSource = env.socketTextStream("localhost", 8899);//streamSource.writeAsText("datas/socket", FileSystem.WriteMode.OVERWRITE).setParallelism(1);DataStreamSource<Tuple2<String, Integer>> streamSource = env.fromElements(Tuple2.of("篮球", 1),Tuple2.of("篮球", 2),Tuple2.of("篮球", 3),Tuple2.of("足球", 3),Tuple2.of("足球", 2),Tuple2.of("足球", 3));// writeAsCsv 只能保存 tuple类型的DataStream流，因为如果不是多列的话，没必要使用什么分隔符streamSource.writeAsCsv("datas/csv", FileSystem.WriteMode.OVERWRITE).setParallelism(1);//5. execute-执行env.execute();}
}

3、连接器Connectors

JDBC Connector

该连接器可以向JDBC 数据库写入数据

JDBC | Apache Flink

<dependency><groupId>org.apache.flink</groupId><artifactId>flink-connector-jdbc_2.11</artifactId><version>${flink.version}</version>
</dependency><!--假如你是连接低版本的，使用5.1.49--><dependency><groupId>mysql</groupId><artifactId>mysql-connector-java</artifactId><version>8.0.25</version></dependency>

案例演示：

将结果读取，写入到MySQL

package com.bigdata.day03;import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.NoArgsConstructor;
import org.apache.flink.connector.jdbc.JdbcConnectionOptions;
import org.apache.flink.connector.jdbc.JdbcExecutionOptions;
import org.apache.flink.connector.jdbc.JdbcSink;
import org.apache.flink.connector.jdbc.JdbcStatementBuilder;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;import java.sql.PreparedStatement;
import java.sql.SQLException;@Data
@AllArgsConstructor
@NoArgsConstructor
class Student{private int id;private String name;private int age;
}
public class JdbcSinkDemo {public static void main(String[] args) throws Exception {StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();DataStreamSource<Student> studentStream = env.fromElements(new Student(1, "jack", 54));JdbcConnectionOptions jdbcConnectionOptions = new JdbcConnectionOptions.JdbcConnectionOptionsBuilder().withUrl("jdbc:mysql://localhost:3306/test1").withDriverName("com.mysql.jdbc.Driver").withUsername("root").withPassword("123456").build();studentStream.addSink(JdbcSink.sink("insert into student values(null,?,?)",new JdbcStatementBuilder<Student>() {@Overridepublic void accept(PreparedStatement preparedStatement, Student student) throws SQLException {preparedStatement.setString(1,student.getName());preparedStatement.setInt(2,student.getAge());}// 假如是流的方式可以设置两条插入一次}, JdbcExecutionOptions.builder().withBatchSize(2).build(),jdbcConnectionOptions));env.execute();}
}

package com.bigdata.day03;import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.NoArgsConstructor;
import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.connector.jdbc.JdbcConnectionOptions;
import org.apache.flink.connector.jdbc.JdbcExecutionOptions;
import org.apache.flink.connector.jdbc.JdbcSink;
import org.apache.flink.connector.jdbc.JdbcStatementBuilder;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;import java.sql.PreparedStatement;
import java.sql.SQLException;@Data
@AllArgsConstructor
@NoArgsConstructor
class Student{private int id;private String name;private int age;
}
public class Demo03 {public static void main(String[] args) throws Exception {//1. env-准备环境StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);DataStreamSource<Student> studentDataStreamSource = env.fromElements(new Student(1, "张三", 19),new Student(2, "lisi", 20),new Student(3, "wangwu", 19));JdbcConnectionOptions jdbcConnectionOptions = new JdbcConnectionOptions.JdbcConnectionOptionsBuilder().withUrl("jdbc:mysql://localhost:3306/kettle").withDriverName("com.mysql.cj.jdbc.Driver").withUsername("root").withPassword("root").build();studentDataStreamSource.addSink(JdbcSink.sink("insert into student values(null,?,?)",new JdbcStatementBuilder<Student>() {@Overridepublic void accept(PreparedStatement preparedStatement, Student student) throws SQLException {preparedStatement.setString(1,student.getName());preparedStatement.setInt(2,student.getAge());}},jdbcConnectionOptions));//2. source-加载数据//3. transformation-数据处理转换//4. sink-数据输出//5. execute-执行env.execute();}
}

运行结果正常：

KafkaConnector

Kafka | Apache Flink

从Kafka的topic1中消费日志数据,并做实时ETL,将状态为success的数据写入到Kafka的topic2中

kafka-topics.sh --bootstrap-server bigdata01:9092 --create --partitions 1 --replication-factor 3 --topic topic1
kafka-topics.sh --bootstrap-server bigdata01:9092 --create --partitions 1 --replication-factor 3 --topic topic2使用控制台当做kafka消息的生产者向kafka中的topic1 发送消息
kafka-console-producer.sh  --bootstrap-server bigdata01:9092 --topic topic1消费kafka中topic2中的数据
kafka-console-consumer.sh  --bootstrap-server bigdata01:9092 --topic topic2操作：
通过黑窗口向topic1中发送消息，含有success字样的消息，会出现在topic2中。

package com.bigdata.day03;import org.apache.flink.api.common.RuntimeExecutionMode;
import org.apache.flink.api.common.functions.FilterFunction;
import org.apache.flink.api.common.serialization.SimpleStringSchema;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaConsumer;
import org.apache.flink.streaming.connectors.kafka.FlinkKafkaProducer;
import org.apache.kafka.clients.producer.KafkaProducer;import java.util.Properties;public class KafkaSinkDemo {// 从topic1中获取数据，放入到topic2中，训练了读和写public static void main(String[] args) throws Exception {//1. env-准备环境StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);//2. source-加载数据Properties properties = new Properties();properties.setProperty("bootstrap.servers", "bigdata01:9092");properties.setProperty("group.id", "g1");FlinkKafkaConsumer<String> kafkaSource = new FlinkKafkaConsumer<String>("topic1",new SimpleStringSchema(),properties);DataStreamSource<String> dataStreamSource = env.addSource(kafkaSource);//3. transformation-数据处理转换SingleOutputStreamOperator<String> filterStream = dataStreamSource.filter(new FilterFunction<String>() {@Overridepublic boolean filter(String s) throws Exception {return s.contains("success");}});//4. sink-数据输出FlinkKafkaProducer kafkaProducer = new FlinkKafkaProducer<String>("topic2",new SimpleStringSchema(),properties);filterStream.addSink(kafkaProducer);//5. execute-执行env.execute();}
}

Flink Kafka Consumer 需要知道如何将 Kafka 中的二进制数据转换为 Java 或者Scala 对象。KafkaDeserializationSchema 允许用户指定这样的 schema，每条 Kafka 中的消息会调用 T deserialize(ConsumerRecord<byte[], byte[]> record) 反序列化。

为了方便使用，Flink 提供了以下几种 schemas：

SimpleStringSchema：按照字符串方式序列化、反序列化

剩余还有 TypeInformationSerializationSchema、JsonDeserializationSchema、AvroDeserializationSchema等。

自定义Sink--模拟jdbcSink的实现

jdbcSink官方已经提供过了，此处仅仅是模拟它的实现，从而学习如何自定义sink

package com.bigdata.day03;import lombok.AllArgsConstructor;
import lombok.Data;
import lombok.NoArgsConstructor;
import org.apache.flink.configuration.Configuration;
import org.apache.flink.connector.jdbc.JdbcConnectionOptions;
import org.apache.flink.connector.jdbc.JdbcExecutionOptions;
import org.apache.flink.connector.jdbc.JdbcSink;
import org.apache.flink.connector.jdbc.JdbcStatementBuilder;
import org.apache.flink.streaming.api.datastream.DataStreamSource;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.sink.RichSinkFunction;import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.PreparedStatement;
import java.sql.SQLException;public class CustomJdbcSinkDemo {@Data@AllArgsConstructor@NoArgsConstructorstatic class Student{private int id;private String name;private int age;}static class MyJdbcSink  extends RichSinkFunction<Student> {Connection conn =null;PreparedStatement ps = null;@Overridepublic void open(Configuration parameters) throws Exception {// 这个里面编写连接数据库的代码Class.forName("com.mysql.jdbc.Driver");conn = DriverManager.getConnection("jdbc:mysql://localhost:3306/test1", "root", "123456");ps = conn.prepareStatement("INSERT INTO `student` (`id`, `name`, `age`) VALUES (null, ?, ?)");}@Overridepublic void close() throws Exception {// 关闭数据库的代码ps.close();conn.close();}@Overridepublic void invoke(Student student, Context context) throws Exception {// 将数据插入到数据库中ps.setString(1,student.getName());ps.setInt(2,student.getAge());ps.execute();}}public static void main(String[] args) throws Exception {StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();DataStreamSource<Student> studentStream = env.fromElements(new Student(1, "马斯克", 51));studentStream.addSink(new MyJdbcSink());env.execute();}
}

查看全文

http://www.lryc.cn/news/491926.html