如何使用Flink SQL按事件时间对流进行排序

Question

我有一个无序的DataStream

public static void main(String[] args) throws Exception {
    StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
    StreamTableEnvironment tableEnv = TableEnvironment.getTableEnvironment(env);

    env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime);
    env.setParallelism(1);

    DataStream<Event> eventStream = env.addSource(new OutOfOrderEventSource())
            .assignTimestampsAndWatermarks(new TimestampsAndWatermarks());

    Table events = tableEnv.fromDataStream(eventStream, "timestamp.rowtime");
    tableEnv.registerTable("events", events);
    Table sorted = tableEnv.sqlQuery("SELECT timestamp FROM events ORDER BY eventTime ASC");
    DataStream<Row> sortedEventStream = tableEnv.toAppendStream(sorted, Row.class);

    sortedEventStream.print();

    env.execute();
}

我得到这个错误：

线程"main"中的异常org.apache.flink.table.api.SqlParserException：SQL解析失败。在第1行，第8列遇到"时间戳From"。

似乎我没有以正确的方式指定事件时间属性，但不清楚出了什么问题。




Answer 1

问题是在我的事件类中使用时间戳作为字段名。将其更改为eventTime（事件时间）就足以让一切正常工作：

public class Sort {
    public static final int OUT_OF_ORDERNESS = 1000;

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

        StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
        StreamTableEnvironment tableEnv = TableEnvironment.getTableEnvironment(env);

        env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime);
        env.setParallelism(1);

        DataStream<Event> eventStream = env.addSource(new OutOfOrderEventSource())
                .assignTimestampsAndWatermarks(new TimestampsAndWatermarks());

        Table events = tableEnv.fromDataStream(eventStream, "eventTime.rowtime");
        tableEnv.registerTable("events", events);
        Table sorted = tableEnv.sqlQuery("SELECT eventTime FROM events ORDER BY eventTime ASC");
        DataStream<Row> sortedEventStream = tableEnv.toAppendStream(sorted, Row.class);

        sortedEventStream.print();

        env.execute();
    }

    public static class Event {
        public Long eventTime;

        Event() {
            this.eventTime = Instant.now().toEpochMilli() + (new Random().nextInt(OUT_OF_ORDERNESS));
        }
    }

    private static class OutOfOrderEventSource implements SourceFunction<Event> {
        private volatile boolean running = true;

        @Override
        public void run(SourceContext<Event> ctx) throws Exception {
            while(running) {
                ctx.collect(new Event());
                Thread.sleep(1);
            }
        }

        @Override
        public void cancel() {
            running = false;
        }
    }

    private static class TimestampsAndWatermarks extends BoundedOutOfOrdernessTimestampExtractor<Event> {
        public TimestampsAndWatermarks() {
            super(Time.milliseconds(OUT_OF_ORDERNESS));
        }

        @Override
        public long extractTimestamp(Event event) {
            return event.eventTime;
        }
    }
}