Flink--Table API 和SQL入门案例

流式表处理的过程可以描述出下图：

 1.先把流转换成动态表   Stream=>  table

2.表经过一定操作，转成结果表   table  （CRUD）=>  table

3.将表转换成流输出      table => Stream

/**
 * @program: flink
 * @description: ${description}
 * @author: Mr.G
 * @create: 2021-09-26 19:21
 **/
package com.ct.day08

import com.ct.day01.SensorSource
import org.apache.flink.streaming.api.scala.StreamExecutionEnvironment
import org.apache.flink.table.api._
import org.apache.flink.api.scala._
import org.apache.flink.table.api.scala._
import org.apache.flink.types.Row


/**
 * @ClassName: ScalarFunction
 * @Description: ${description}
 * @Author Mr.G
 * @Date 2021/9/26
 * @Version 1.0
 *
 */
object TableApiExample {

  def main(args: Array[String]): Unit = {

    val  env = StreamExecutionEnvironment.getExecutionEnvironment

    env.setParallelism(1)

    val settings = EnvironmentSettings
      .newInstance()
      .useBlinkPlanner()
      .inStreamingMode()
      .build()

    //得到表处理运行环境
    val tEnv = StreamTableEnvironment.create(env,settings)

    val stream = env.addSource(new SensorSource)


    // stream => table

    val table: Table = tEnv.fromDataStream(stream,'id,'timestamp as 'ts,'temperature)


    // TableApi 形式
    // table   CRUD   => table
    val rstable: Table = table
      .filter("id ='sensor_1'")
      .select('id)
    // table => DStream
    tEnv.toAppendStream[Row](rstable)
//      .print()


    // SQL 形式

    //先注册成一个表(视图）    Stream => table
    tEnv.createTemporaryView("sensor",stream,'id,'timestamp as 'ts,'temperature)


    //table   CRUD  => table
    tEnv.sqlQuery(" select id from sensor where id = 'sensor_1'")
      //table => Stream
      .toAppendStream[Row]
     // .toRetractStream[Row]
        .print()

    env.execute()

  }

}