【云星数据---Apache Flink实战系列(精品版)】：Apache Flink高级特性与高级应用015-Flink中广播变量和分布式缓存001

1.flink中的广播变量

flink支持将变量广播到worker上，以供程序运算使用。

执行程序

package code.book.batch.sinksource.scalaimport java.utilimport org.apache.flink.api.common.functions.RichMapFunctionimport org.apache.flink.api.scala.{DataSet, ExecutionEnvironment, _}import org.apache.flink.configuration.Configurationobject BroadcastVariables001 {  def main(args: Array[String]): Unit = {    val env = ExecutionEnvironment.getExecutionEnvironment    //1.准备工人数据（用于map）    case class Worker(name: String, salaryPerMonth: Double)    val workers: DataSet[Worker] = env.fromElements(      Worker("zhagnsan", 1356.67),      Worker("lisi", 1476.67)    )    //2准备统计数据(用于广播，通过withBroadcastSet进行广播)    case class Count(name: String, month: Int)    val counts: DataSet[Count] = env.fromElements(      Count("zhagnsan", 4),      Count("lisi", 5)    )    //3.使用map数据和广播数据进行计算    workers.map(new RichMapFunction[Worker, Worker] {      private var cwork: util.List[Count] = null      override def open(parameters: Configuration): Unit = {        super.open(parameters)        // 3.1 访问广播数据        cwork = getRuntimeContext.getBroadcastVariable[Count]("countWorkInfo")      }      override def map(w: Worker): Worker = {        //3.2解析广播数据        var i = 0        while (i < cwork.size()) {          val c = cwork.get(i)          i += 1          if (c.name.equalsIgnoreCase(w.name)) {            //有相应的信息的返回值            return Worker(w.name, w.salaryPerMonth * c.month)          }        }        //无相应的信息的返回值        Worker("###", 0)      }    }).withBroadcastSet(counts, "countWorkInfo").print()  }}

执行效果

Worker(zhagnsan,5426.68)Worker(lisi,7383.35)