flink分别使用FilterMap和ProcessFunction实现去重逻辑

背景

在日常的工作中，对数据去重是一件很常见的操作，比如我们只需要保留重复记录的第一条，而忽略掉后续重复的记录，达到去重的效果，本文就使用flink的FilterMap和ProcessFunction来实现去重逻辑

FilterMap和ProcessFunction去重实现

filterMap实现去重

java 复制代码

public class DuplicateRichFlatMap extends RichFlatMapFunction<WikipediaEditEvent, WikipediaEditEvent> {

    ValueState<Boolean> duplicateInput;

    @Override
    public void open(Configuration parameters) throws Exception {
        duplicateInput = getRuntimeContext().getState(new ValueStateDescriptor<Boolean>("duplicate", Types.BOOLEAN));
    }

    @Override
    public void flatMap(WikipediaEditEvent in, Collector<WikipediaEditEvent> collector) throws Exception {
        if (duplicateInput.value() == null) {
            collector.collect(in);
            duplicateInput.update(true);
        }
    }

}

这里实现的关键就是有一个key-value的flink状态

ProcessFunction去重

java 复制代码

public class DupliacateProcessFunction extends KeyedProcessFunction<String, WikipediaEditEvent, WikipediaEditEvent> {


    ValueState<Boolean> duplicateInput;

    @Override
    public void open(Configuration parameters) throws Exception {
        ValueStateDescriptor stateDescriptor = new ValueStateDescriptor<Boolean>("previousInput", Types.BOOLEAN);
        // 状态ttl超时时间设置
        StateTtlConfig ttlConfig =
                StateTtlConfig.newBuilder(Time.days(1)).setUpdateType(StateTtlConfig.UpdateType.OnCreateAndWrite)
                        .setStateVisibility(StateTtlConfig.StateVisibility.NeverReturnExpired)
                        // check 10 keys for every state access
                        .cleanupIncrementally(100, false).build();
        stateDescriptor.enableTimeToLive(ttlConfig);
        duplicateInput = getRuntimeContext().getState(stateDescriptor);
    }

    @Override
    public void processElement(WikipediaEditEvent in, Context context, Collector<WikipediaEditEvent> collector)
            throws Exception {
        if (duplicateInput.value() == null) {
            collector.collect(in);
            duplicateInput.update(true);
        }
    }
}

这里的关键代码也是拥有一个key-value的状态

触发计算的job代码如下

java 复制代码

public class DuplicateJob {

    public static void main(String[] args) throws Exception {

        StreamExecutionEnvironment see = StreamExecutionEnvironment.getExecutionEnvironment();
        see.setStreamTimeCharacteristic(TimeCharacteristic.ProcessingTime);

        DataStream<WikipediaEditEvent> edits = see.addSource(new RandomStringSource());

        KeyedStream<WikipediaEditEvent, String> keyedEdits = edits.keyBy(new KeySelector<WikipediaEditEvent, String>() {
            @Override
            public String getKey(WikipediaEditEvent event) {
                return event.getUser();
            }
        });

        // 通过RichFlatMap实现去重
        DataStream<WikipediaEditEvent> result = keyedEdits.flatMap(new DuplicateRichFlatMap());
        // 通过ProcessFunction实现去重
//        DataStream<WikipediaEditEvent> result = keyedEdits.process(new DupliacateProcessFunction());

        result.print();
        see.execute();

    }
}