更新模式

对于stream类型的Table数据,需要标记是由于INSERT,UPDATE,DELETE中哪种操作更新的数据,在Table API中通过Update Modes指定数据更新类型,通过指定不同的Update Modes确定哪种更新操作的数据与外部系统进行交互

.connect(...)
.inAppendMode()		//交互INSERT操作更新数据
.inUpsertMode()		//INSERT,UPDATE,DELETE操作更新数据
.inRetractMode()	//交互INSERT和DELETE操作更新数据

应用实例

import org.apache.flink.api.scala.typeutils.Types
import org.apache.flink.streaming.api.scala._
import org.apache.flink.table.api.scala._
import org.apache.flink.table.descriptors.{Json, Kafka, Rowtime, Schema}

object ConnectorDemo {
  def main(args: Array[String]): Unit = {
    val env = StreamExecutionEnvironment.getExecutionEnvironment
    val tabEnv = StreamTableEnvironment.create(env)
    tabEnv.connect(
      new Kafka()
        .version("0.10")
        .topic("test")
        .property("zookeeper.connect","note01:2181,note02:2181,note03:2181")
        .property("bootstrap.servers","note01:9092,note02:9092,note03:9092")
    )
    //指定Table Format信息
      .withFormat(
        new Json()
          .failOnMissingField(true)
          .jsonSchema(
            """
              |{
              |type:'object',
              |properties:{
              |id:{
              |type:'number'
              |},
              |name:{
              |type:'string'
              |},
              |timestamp:{
              |type:'string'      
              |format:'date-time'
              |}
              |}
              |}
              |""".stripMargin)
      ).withSchema(
      new Schema()
        .field("id",Types.INT)
        .field("name",Types.STRING)
        .field("rowtime",Types.SQL_TIMESTAMP)
        .rowtime(
          new Rowtime()
            .timestampsFromField("timestamp")
            .watermarksPeriodicBounded(60000) //1 minutes
        )
    )
    //指定数据更新模式
      .inAppendMode()
      .registerTableSource("kafkaTable")
  }
}