「Flink」使用Java lambda表达式实现Flink WordCount
本篇我们将使用Java语言来实现Flink的单词统计。
代码开发
环境准备
导入Flink 1.9 pom依赖
<dependencies>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-java</artifactId>
<version>1.9.0</version>
</dependency>
<dependency>
<groupId>org.apache.flink</groupId>
<artifactId>flink-streaming-java_2.11</artifactId>
<version>1.9.0</version>
</dependency>
<dependency>
<groupId>org.apache.commons</groupId>
<artifactId>commons-lang3</artifactId>
<version>3.7</version>
</dependency>
</dependencies>
构建Flink流处理环境
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
自定义source
每秒生成一行文本
DataStreamSource<String> wordLineDS = env.addSource(new RichSourceFunction<String>() {
private boolean isCanal = false;
private String[] words = {
"important oracle jdk license update",
"the oracle jdk license has changed for releases starting april 16 2019",
"the new oracle technology network license agreement for oracle java se is substantially different from prior oracle jdk licenses the new license permits certain uses such as ",
"personal use and development use at no cost but other uses authorized under prior oracle jdk licenses may no longer be available please review the terms carefully before ",
"downloading and using this product an faq is available here ",
"commercial license and support is available with a low cost java se subscription",
"oracle also provides the latest openjdk release under the open source gpl license at jdk java net"
};
@Override
public void run(SourceContext<String> ctx) throws Exception {
// 每秒发送一行文本
while (!isCanal) {
int randomIndex = RandomUtils.nextInt(0, words.length);
ctx.collect(words[randomIndex]);
Thread.sleep(1000);
}
}
@Override
public void cancel() {
isCanal = true;
}
});
单词计算
// 3. 单词统计
// 3.1 将文本行切分成一个个的单词
SingleOutputStreamOperator<String> wordsDS = wordLineDS.flatMap((String line, Collector<String> ctx) -> {
// 切分单词
Arrays.stream(line.split(" ")).forEach(word -> {
ctx.collect(word);
});
}).returns(Types.STRING); //3.2 将单词转换为一个个的元组
SingleOutputStreamOperator<Tuple2<String, Integer>> tupleDS = wordsDS
.map(word -> Tuple2.of(word, 1))
.returns(Types.TUPLE(Types.STRING, Types.INT)); // 3.3 按照单词进行分组
KeyedStream<Tuple2<String, Integer>, String> keyedDS = tupleDS.keyBy(tuple -> tuple.f0); // 3.4 对每组单词数量进行累加
SingleOutputStreamOperator<Tuple2<String, Integer>> resultDS = keyedDS
.timeWindow(Time.seconds(3))
.reduce((t1, t2) -> Tuple2.of(t1.f0, t1.f1 + t2.f1)); resultDS.print();
参考代码
public class WordCount {
public static void main(String[] args) throws Exception {
// 1. 构建Flink流式初始化环境
StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
// 2. 自定义source - 每秒发送一行文本
DataStreamSource<String> wordLineDS = env.addSource(new RichSourceFunction<String>() {
private boolean isCanal = false;
private String[] words = {
"important oracle jdk license update",
"the oracle jdk license has changed for releases starting april 16 2019",
"the new oracle technology network license agreement for oracle java se is substantially different from prior oracle jdk licenses the new license permits certain uses such as ",
"personal use and development use at no cost but other uses authorized under prior oracle jdk licenses may no longer be available please review the terms carefully before ",
"downloading and using this product an faq is available here ",
"commercial license and support is available with a low cost java se subscription",
"oracle also provides the latest openjdk release under the open source gpl license at jdk java net"
};
@Override
public void run(SourceContext<String> ctx) throws Exception {
// 每秒发送一行文本
while (!isCanal) {
int randomIndex = RandomUtils.nextInt(0, words.length);
ctx.collect(words[randomIndex]);
Thread.sleep(1000);
}
}
@Override
public void cancel() {
isCanal = true;
}
});
// 3. 单词统计
// 3.1 将文本行切分成一个个的单词
SingleOutputStreamOperator<String> wordsDS = wordLineDS.flatMap((String line, Collector<String> ctx) -> {
// 切分单词
Arrays.stream(line.split(" ")).forEach(word -> {
ctx.collect(word);
});
}).returns(Types.STRING);
//3.2 将单词转换为一个个的元组
SingleOutputStreamOperator<Tuple2<String, Integer>> tupleDS = wordsDS
.map(word -> Tuple2.of(word, 1))
.returns(Types.TUPLE(Types.STRING, Types.INT));
// 3.3 按照单词进行分组
KeyedStream<Tuple2<String, Integer>, String> keyedDS = tupleDS.keyBy(tuple -> tuple.f0);
// 3.4 对每组单词数量进行累加
SingleOutputStreamOperator<Tuple2<String, Integer>> resultDS = keyedDS
.timeWindow(Time.seconds(3))
.reduce((t1, t2) -> Tuple2.of(t1.f0, t1.f1 + t2.f1));
resultDS.print();
env.execute("app");
}
}
Flink对Java Lambda表达式支持情况
Flink支持Java API所有操作符使用Lambda表达式。但是,但Lambda表达式使用Java泛型时,就需要声明类型信息。
我们来看下上述的这段代码:
SingleOutputStreamOperator<String> wordsDS = wordLineDS.flatMap((String line, Collector<String> ctx) -> {
// 切分单词
Arrays.stream(line.split(" ")).forEach(word -> {
ctx.collect(word);
});
}).returns(Types.STRING);
之所以这里将所有的类型信息,因为Flink无法正确自动推断出来Collector中带的泛型。我们来看一下FlatMapFuntion的源代码
@Public
@FunctionalInterface
public interface FlatMapFunction<T, O> extends Function, Serializable {
/**
* The core method of the FlatMapFunction. Takes an element from the input data set and transforms
* it into zero, one, or more elements.
*
* @param value The input value.
* @param out The collector for returning result values.
*
* @throws Exception This method may throw exceptions. Throwing an exception will cause the operation
* to fail and may trigger recovery.
*/
void flatMap(T value, Collector<O> out) throws Exception;
}
我们发现 flatMap的第二个参数是Collector<O>,是一个带参数的泛型。Java编译器编译该代码时会进行参数类型擦除,所以Java编译器会变成成:
void flatMap(T value, Collector out)
这种情况,Flink将无法自动推断类型信息。如果我们没有显示地提供类型信息,将会出现以下错误:
org.apache.flink.api.common.functions.InvalidTypesException: The generic type parameters of 'Collector' are missing.
In many cases lambda methods don't provide enough information for automatic type extraction when Java generics are involved.
An easy workaround is to use an (anonymous) class instead that implements the 'org.apache.flink.api.common.functions.FlatMapFunction' interface.
Otherwise the type has to be specified explicitly using type information.
这种情况下,必须要显示指定类型信息,否则输出将返回值视为Object类型,这将导致Flink无法正确序列化。
所以,我们需要显示地指定Lambda表达式的参数类型信息,并通过returns方法显示指定输出的类型信息
我们再看一段代码:
SingleOutputStreamOperator<Tuple2<String, Integer>> tupleDS = wordsDS
.map(word -> Tuple2.of(word, 1))
.returns(Types.TUPLE(Types.STRING, Types.INT));
为什么map后面也需要指定类型呢?
因为此处map返回的是Tuple2类型,Tuple2是带有泛型参数,在编译的时候同样会被查出泛型参数信息,导致Flink无法正确推断。
更多关于对Java Lambda表达式的支持请参考官网:https://ci.apache.org/projects/flink/flink-docs-release-1.9/dev/java_lambdas.html
「Flink」使用Java lambda表达式实现Flink WordCount的更多相关文章
- Java Lambda表达式初探
Java Lambda表达式初探 前言 本文受启发于Trisha Gee在JavaOne 2016的主题演讲Refactoring to Java 8. Java 8已经发行两年多,但很多人仍然在使用 ...
- Java Lambda表达式入门
Java Lambda表达式入门 http://blog.csdn.net/renfufei/article/details/24600507 Java 8十个lambda表达式案例 http://w ...
- Java Lambda表达式入门[转]
原文链接: Start Using Java Lambda Expressions http://blog.csdn.net/renfufei/article/details/24600507 下载示 ...
- Java Lambda表达式教程与示例
Lambda表达式是Java 8中引入的一个新特性.一个lambda表达式是一个匿名函数,而且这个函数没有名称且不属于任何类.lambda表达式的概念最初是在LISP编程语言中引入的. Java La ...
- Java Lambda表达式forEach无法跳出循环的解决思路
Java Lambda表达式forEach无法跳出循环的解决思路 如果你使用过forEach方法来遍历集合,你会发现在lambda表达式中的return并不会终止循环,这是由于lambda的底层实现导 ...
- java lambda表达式学习笔记
lambda是函数式编程(FP,functional program),在java8中引入,而C#很早之前就有了.在java中lambda表达式是'->',在C#中是‘=>’. 杜甫说:射 ...
- 《Java基础知识》Java Lambda表达式
接触Lambda表达式的时候,第一感觉就是,这个是啥?我居然看不懂,于是开始寻找资料,必须弄懂它. 先来看一个案例: @FunctionalInterface public interface MyL ...
- Java lambda 表达式常用示例
实体类 package com.lkb.java_lambda.dto; import lombok.Data; /** * @program: java_lambda * @description: ...
- Java lambda 表达式详解(JDK 8 新特性)
什么是 lambda 表达式 lambda 表达式(拉姆达表达式)是 JAVA 8 中提供的一种新的特性,它使 Java 也能进行简单的"函数式编程". lambda 表达式的本质 ...
随机推荐
- 正斜杠(" / ")和反斜杠(" \ ")的区别
反斜杠“\”是电脑出现了之后为了表示程序设计里的特殊含义才发明的专用标点.所以除了程序设计领域外,任何地方都不应该使用反斜杠. 如何区分正反斜杠 英语:"/" 英文是forward ...
- python 安装虚拟环境virtualenv
1.sudo apt install virtualenv 安装失败 2.sudo apt-get update 更新失败 提示: E: 仓库 “http://mirrors.aliyun.com/u ...
- [计算几何+图论]doge
题意 在平面直角坐标系上,你有一只doge在原点处.doge被绳子拴住了,绳子不会打结,没有弹性(但很柔软),并且长度为L.平面上有一些目标,因此你的doge会按照顺序去捡起它们,但是doge只能走直 ...
- Docker底层架构之基础架构
Docker 采用了 C/S架构,包括客户端和服务端. Docker daemon 作为服务端接受来自客户 的请求,并处理这些请求(创建.运行.分发容器). 客户端和服务端既可以运行在一个机器上,也可 ...
- laravel 初学路由简单介绍
文档中的路由详细演示[初学laravel]对应laravel 的框架目录:routes/web.php 路由的格式一:Route::get($uri,$callback); 1.简单的浏览器输出 Ro ...
- mysql-cluster集群搭建步骤
1.从官网下载mysql-cluster安装包: https://dev.mysql.com/downloads/cluster/ 2.解压安装包 #上传到服务器目录/usr/softwares并解压 ...
- 前端入门nginx
一.nginx是什么 NGINX is a free, open-source, high-performance HTTP server and reverse proxy, as well as ...
- 前端 JS/TS 调用 ASP.NET Core gRPC-Web
前言 在上两篇文章中,介绍了ASP.NET Core 中的 gRPC-Web 实现 和 在 Blazor WebAssembly 中使用 gRPC-Web,实现了 Blazor WebAssembly ...
- Arduino系列之DHT11模块采集数据(一)
下面我将介绍DHT11模块的相关用法 DHT11数字传感器概述:是一款含有已校准数字信号输出的温湿度复合传感器 .它应用专用的数字模块采集技术和温湿度传感技术 ,确保产品具有极 高的可靠性与卓越的长期 ...
- RabbitMQ下载与安装
RabbitMQ下载与安装 先跟大家科普一下MQ和RabbitMQ MQ简介 MQ全称为Message Queue ,即消息队列 应用场景: 1.任务异步处理. 将不需要同步处理的并且耗时长的操作由消 ...