4. flink通过构造函数将参数传统给算子《深入理解flink系列》

1.《深入理解Flink核心设计与实践原理》一书作者
2.GitHub 热门项目 fink-boot（800+） 开发者，致力于flink与spring生态集成 3.前上市top咨询公司技术经理，现大厂资深开发
4.证书:软件设计师，Apache kylin管理员
5.方向:java服务端开发，分布式开发，实时计算，大数据开发工程师

通过构造函数参数传递给算子

Flink提供了三种方式来简化将参数传递给用户定义的函数这一过程：使用构造函数或withParameters(Configuration)方法或者ExecutionConfig接口都可以将参数传递给函数。

1 通过构造函数

import org.apache.flink.api.common.functions.FilterFunction;
import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;

public class ConstructorTemplate {

    public static void main(String[] args) throws Exception {
        final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
        env.setParallelism(2);
       
        //生成一个1到20的数字序列的数据流。
        DataStream<Long> dataStream = env.generateSequence(1,20);

        //将ParamBean对象的参数传入Filter操作实例中
        DataStream<Long> resultStream=dataStream.filter(new FilterConsrtucted(new ParamBean("intsmaze",10)))；

        resultStream.print("constructed stream is ");
        env.execute("ParamTemplate intsmaze");
    }
    
    //用户定义的Filter函数实现了FilterFunction接口
    private static class FilterConsrtucted implements FilterFunction<Long> {

        private final ParamBean paramBean;
        //Filter函数的构造函数参数为ParamBean对象，用于接收传入的参数值
        public FilterConsrtucted(ParamBean paramBean) {
            this.paramBean = paramBean;
        }

        @Override
        public boolean filter(Long value) {
            return value > paramBean.getFlag();
        }
    }
}

public class ParamBean implements Serializable {

    private String name;

    private int flag;

    public ParamBean(String name, int flag) {
        this.name = name;
        this.flag = flag;
    }
    ......get,set方法
}

如下输出结果:

constructed stream is :2> 12
constructed stream is :1> 11
constructed stream is :2> 14
constructed stream is :1> 13
constructed stream is :2> 16
constructed stream is :1> 15
constructed stream is :2> 18
constructed stream is :1> 17
constructed stream is :2> 20
constructed stream is :1> 19

注意:如果向用户定义的函数传递的参数不是基本类型，是自定义的类型，那么需要该类型实现Serializable接口，否则运行程序会报以下异常。

Exception in thread "main" org.apache.flink.api.common.InvalidProgramException: The implementation of the RichFilterFunction is not serializable. The object probably contains or references non serializable fields.