Flink原理与实践

初识Flink

环境搭建和部署

入门案例

1，统计词出现格式
1.1，使用Lambda表达式

public class WordCount01 {
	public static void main(String[] args) {
		// 设置用户名在Linux下运行，指定HADOOP用户权限
		System.setProperty("HADOOP_USER_NAME","root");
		//env
		StreamExecutionEnvironment env = StreamExecutionEnvironment.getEnvironment();
		env.setRuntimeMode(RuntimeExecutionMode.AUTOMATIC);//自动匹配，流，批
		
		//source
		DataStream<String> lines= env.fromElements("welcome to flink world", "welcome to flink", "welcome flink");
		//transformation--etl
		// 切割
		DataStream<String>  words= words.flatMap(String word, Collector<String> out->{
		Arrays.stream(word.split(" ")).foreach(out:collect)
		}).returns(Type.STRING);
		// 标记
		DataStream<Tuple2<String,Integer>> wordAndOne = words.map((String value) -> Tuple2.of(value, 1))
    	.returns(Types.TUPLE(Types.STRING, Types.INT));
	    // 分组
	    KeyedStream<Tuple2<String,Integer>,String> grouped = wordAndOne.keyBy((KeySelector<Tuple2<String,Integer>,String>)t->t.f0);
	    // 统计
	    SingleOutputStreamOperator<String> result = grouped.sum(1);
		//sink
		//输出
		env.print();
		//启动并执行
		env.execute("word count");
	}
}