Flink有状态计算入门示例

需求：从socket接收字符串，按空格分隔成单词，统计单词出现的次数。

直接看代码吧。

主函数：

FlinkStatefulCalcTest.java

package com.ccclubs.flink;

import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.util.Collector;

/**
 * @author xianghu.wang
 * @time 2019/4/11
 * @description
 */
public class FlinkStatefulCalcTest {
    public static void main(String[] args) throws Exception {
        // 创建运行环境
        final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();

        // 为方便输出，设置并行度为1
        env.setParallelism(1);

        // 从socket接收数据
        DataStream words = env.socketTextStream("localhost", 9999, "\n");

        // 把字符串分隔成单词及其个数的键值对（未聚合）
        words.flatMap((String line, Collector out) -> {
            for (String s : line.split(" ")) {
                out.collect(new WordCount(s, 1));
            }
        }).returns(WordCount.class)
                // 按key分组，维护state时可认为只有一个key
                .keyBy("word")
                // 开始计算
                .flatMap(new WordCountFunction())
                // 打印在控制台
                .print();

        // print the results with a single thread, rather than in parallel
        env.execute("stateful Window WordCount");
    }

}

单词计数实体类：

WordCount.java

package com.ccclubs.flink;

/**
 * @author xianghu.wang
 * @time 2019/4/11
 * @description
 */
public class WordCount {
    /**
     * 单词
     **/
    private String word;
    /**
     * 单词个数
     **/
    private Integer count;

    public WordCount() {
    }

    public WordCount(String word, Integer count) {
        this.word = word;
        this.count = count;
    }

    public String getWord() {
        return word;
    }

    public void setWord(String word) {
        this.word = word;
    }

    public Integer getCount() {
        return count;
    }

    public void setCount(Integer count) {
        this.count = count;
    }

    @Override
    public String toString() {
        return "WordCount{" +
                "word='" + word + '\'' +
                ", count=" + count +
                '}';
    }
}

状态类：

WordCountState.java

package com.ccclubs.flink;

import java.io.Serializable;

/**
 * @author xianghu.wang
 * @time 2019/4/11
 * @description 单词统计状态类
 */
public class WordCountState implements Serializable {
    /**单词**/
    private String word;
    /**单词个数**/
    private Integer count;

    public String getWord() {
        return word;
    }

    public WordCountState(String word, Integer count) {
        this.word = word;
        this.count = count;
    }

    public void setWord(String word) {
        this.word = word;
    }

    public Integer getCount() {
        return count;
    }

    public void setCount(Integer count) {
        this.count = count;
    }
}

重点！！！！ 计算函数

WordCountFunction.java

package com.ccclubs.flink;

import org.apache.flink.api.common.functions.RichFlatMapFunction;
import org.apache.flink.api.common.state.ValueState;
import org.apache.flink.api.common.state.ValueStateDescriptor;
import org.apache.flink.api.common.typeinfo.TypeHint;
import org.apache.flink.api.common.typeinfo.TypeInformation;
import org.apache.flink.configuration.Configuration;
import org.apache.flink.util.Collector;

/**
 * @author xianghu.wang
 * @time 2019/4/11
 * @description 有状态结算，RichFlatMapFunction的泛型分别为
 * 入参和出参的数据类型（根据实际需求来，这里都是WordCount是巧合）
 */
public class WordCountFunction extends RichFlatMapFunction {
    /**
     * 单词统计状态
     **/
    private transient ValueState countState;

    /**
     * 有状态计算
     *
     * @param in
     * @param out
     * @throws Exception
     */
    @Override
    public void flatMap(WordCount in, Collector out) throws Exception {
        // access the state in
        WordCountState lastState = countState.value();
        // 第一次进入计算,更新state,输入原样返回
        if (lastState == null) {
            // 初始化state
            lastState = new WordCountState(in.getWord(), in.getCount());

            // 进来的数据原样返回
            out.collect(in);
            // 更新state
            countState.update(lastState);
        } else {
            // count 累加
            lastState.setCount(lastState.getCount() + in.getCount());
            out.collect(new WordCount(in.getWord(), lastState.getCount()));
            // 更新state
            countState.update(lastState);
        }
    }

    /**
     * 初始化ValueState(未指定默认值，官方推荐自己处理)
     *
     * @param config
     */
    @Override
    public void open(Configuration config) {
        ValueStateDescriptor descriptor =
                new ValueStateDescriptor<>(
                        // the state name
                        "countState",
                        // type information
                        TypeInformation.of(new TypeHint() {
                        }));
        countState = getRuntimeContext().getState(descriptor);
    }
}

运行结果：

 

 

参考：

https://ci.apache.org/projects/flink/flink-docs-release-1.8/dev/stream/state/state.html

注：转载请注明出处