flink-table-sql-demo1

一.背景

flink 这个东西，后面会尝试走纯 SQL 统计路线，这个阿里和华为都搞了一套，这里就简单记录下测试效果。

用SQL统计用户点击数，每隔5秒统计一次。暂时去掉了复杂逻辑。

二.直接看代码

// lombok 插件，这里主要写一个简单的数据产生的对象
// 表是时间，用户,以及商品3个字段
@Data
@ToString
public class UserInfo implements Serializable {
    private Timestamp pTime;
    private String userId;
    private String itemId;

    public UserInfo() {
    }

    public UserInfo(String userId, String itemId) {
        this.userId = userId;
        this.itemId = itemId;
        this.pTime = new Timestamp(System.currentTimeMillis());
    }
}

/**
 * 模拟数据产生，没隔1秒发送一个数据
 * @author <a href="mailto:huoguo@2dfire.com">火锅</a>
 * @time 2019/2/22
 */
public class UserDataSource implements SourceFunction<UserInfo> {
    static String[] items = {"i-1", "i-2", "i-3"};
    static String[] users = {"a", "b", "c"};
    @Override
    public void run(SourceContext sc) throws Exception {
        while (true) {
            TimeUnit.SECONDS.sleep(1);
            int m = (int) (System.currentTimeMillis() % 3);
            sc.collect(new UserInfo(users[m], items[m]));
        }
    }
    @Override
    public void cancel() {
        System.out.println("cancel to do ...");
    }
}

/**
 * 这就主函数，负责统计，引用是 java 的，别引错了
 */
public class UserApp {
    public static void main(String[] args) throws Exception {

        StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
        env.setParallelism(1);
        env.setStreamTimeCharacteristic(TimeCharacteristic.EventTime);
        DataStream<UserInfo> userInfoDataStream = env.addSource(new UserDataSource());

        DataStream<UserInfo> timedData = userInfoDataStream.assignTimestampsAndWatermarks(new AscendingTimestampExtractor<UserInfo>() {
            @Override
            public long extractAscendingTimestamp(UserInfo element) {
                return element.getPTime().getTime();
            }
        });
        StreamTableEnvironment tableEnv = TableEnvironment.getTableEnvironment(env);
        // pTime.rowtime  = pTime as rowTime(proctime)
        tableEnv.registerDataStream("test", timedData, "userId,itemId,pTime.rowtime");
        Table result = tableEnv.sqlQuery("SELECT userId,TUMBLE_END(pTime, INTERVAL '5' SECOND) as pTime,count(1) as cnt FROM  test" +
                " GROUP BY TUMBLE(pTime, INTERVAL '5' SECOND),userId ");
        // deal with (Tuple2<Boolean, Row> value) -> out.collect(row)
        SingleOutputStreamOperator allClick = tableEnv.toRetractStream(result, Row.class)
                .flatMap((Tuple2<Boolean, Row> value, Collector<Row> out) -> {
                    out.collect(value.f1);
                }).returns(Row.class);
        // add sink or print
        allClick.print();
        env.execute("test");
    }

}

结果：

c,2019-03-06 13:55:20.0,4

a,2019-03-06 13:55:20.0,1

--- 手动分开好看

c,2019-03-06 13:55:25.0,2

b,2019-03-06 13:55:25.0,2

a,2019-03-06 13:55:25.0,1

---

a,2019-03-06 13:55:30.0,2

c,2019-03-06 13:55:30.0,2

b,2019-03-06 13:55:30.0,1

小结：

1.demo很简单，仅为了测试使用

2.具体的原理，和很多东西 有时间再写吧