AggregateFunction结合自定义触发器实现点击率计算

背景：

接上一篇文章，ProcessWindowFunction 结合自定义触发器会有状态过大的问题，本文就使用AggregateFunction结合自定义触发器来实现，这样就不会导致状态过大的问题了

AggregateFunction结合自定义触发器实现

在这里插入图片描述
flink对于每个窗口只需要维护一个状态：不像ProcessWindowFunction那样需要把窗口内收到的所有消息都作为状态存储起来

在这里插入图片描述
完整代码参见：

package wikiedits.func;import java.text.SimpleDateFormat;
import java.util.Date;import org.apache.flink.api.common.functions.AggregateFunction;
import org.apache.flink.api.java.tuple.Tuple2;
import org.apache.flink.runtime.state.filesystem.FsStateBackend;
import org.apache.flink.streaming.api.CheckpointingMode;
import org.apache.flink.streaming.api.TimeCharacteristic;
import org.apache.flink.streaming.api.datastream.DataStream;
import org.apache.flink.streaming.api.datastream.SingleOutputStreamOperator;
import org.apache.flink.streaming.api.environment.StreamExecutionEnvironment;
import org.apache.flink.streaming.api.functions.source.SourceFunction;
import org.apache.flink.streaming.api.windowing.time.Time;
import org.apache.flink.streaming.api.windowing.triggers.ContinuousProcessingTimeTrigger;public class AggregateFunctionAndTiggerDemo {public static void main(String[] args) throws Exception {final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();// 使用处理时间env.setStreamTimeCharacteristic(TimeCharacteristic.ProcessingTime);env.enableCheckpointing(60000, CheckpointingMode.EXACTLY_ONCE);env.setStateBackend(new FsStateBackend("file:///D:/tmp/flink/checkpoint/aggregatetrigger"));// 并行度为1env.setParallelism(1);// 设置数据源，一共三个元素DataStream<Tuple2<String, Integer>> dataStream = env.addSource(new SourceFunction<Tuple2<String, Integer>>() {@Overridepublic void run(SourceContext<Tuple2<String, Integer>> ctx) throws Exception {int xxxNum = 0;int yyyNum = 0;for (int i = 1; i < Integer.MAX_VALUE; i++) {// 只有XXX和YYY两种nameString name = (0 == i % 2) ? "XXX" : "YYY";// 更新aaa和bbb元素的总数if (0 == i % 2) {xxxNum++;} else {yyyNum++;}// 使用当前时间作为时间戳long timeStamp = System.currentTimeMillis();// 将数据和时间戳打印出来，用来验证数据if (xxxNum % 2000 == 0) {System.out.println(String.format("source，%s, %s,    XXX total : %d,    YYY total : %d\n", name,time(timeStamp), xxxNum, yyyNum));}// 发射一个元素，并且戴上了时间戳ctx.collectWithTimestamp(new Tuple2<String, Integer>(name, 1), timeStamp);// 每发射一次就延时1秒Thread.sleep(1);}}@Overridepublic void cancel() {}});// 将数据用5秒的滚动窗口做划分，再用ProcessWindowFunctionSingleOutputStreamOperator<Tuple2<String, Integer>> mainDataStream = dataStream// 以Tuple2的f0字段作为key，本例中实际上key只有aaa和bbb两种.keyBy(value -> value.f0)// 5秒一次的滚动窗口.timeWindow(Time.minutes(5))// 10s触发一次计算，更新统计结果.trigger(ContinuousProcessingTimeTrigger.of(Time.seconds(10)))// 统计每个key当前窗口内的元素数量，然后把key、数量、窗口起止时间整理成字符串发送给下游算子.aggregate(new AggregateFunction<Tuple2<String, Integer>, Tuple2<String, Integer>, Tuple2<String, Integer>>() {// 1、初始值// 定义累加器初始值@Overridepublic Tuple2<String, Integer> createAccumulator() {return new Tuple2<String, Integer>("", 0);}// 2、累加// 定义累加器如何基于输入数据进行累加@Overridepublic Tuple2<String, Integer> add(Tuple2<String, Integer> value,Tuple2<String, Integer> accumulator) {accumulator.f0 = value.f0;accumulator.f1 += value.f1;return accumulator;}// 3、合并// 定义累加器如何和State中的累加器进行合并@Overridepublic Tuple2<String, Integer> merge(Tuple2<String, Integer> acc1,Tuple2<String, Integer> acc2) {acc1.f1 += acc2.f1;return acc1;}// 4、输出// 定义如何输出数据@Overridepublic Tuple2<String, Integer> getResult(Tuple2<String, Integer> accumulator) {return accumulator;}});// 打印结果，通过分析打印信息，检查ProcessWindowFunction中可以处理所有key的整个窗口的数据mainDataStream.print();env.execute("processfunction demo : processwindowfunction");}public static String time(long timeStamp) {return new SimpleDateFormat("yyyy-MM-dd hh:mm:ss").format(new Date(timeStamp));}}