【sparkStreaming】将DStream保存在MySQL

package SparkDemo

import java.sql.{Connection, DriverManager, PreparedStatement}

import org.apache.spark.SparkConf
import org.apache.spark.streaming.{Seconds, StreamingContext}

object DStreamToMySQL {
    //定义更新函数
    def updateFunc(newValues : Seq[Int],state :Option[Int]):Option[Int] = {

        val currentCount = newValues.foldLeft(0)(_+_)
        val  previousCount = state.getOrElse(0)
        Some(currentCount+previousCount)
    }
    def main(args : Array[String]): Unit ={
        //建立SparkStream
        val conf = new SparkConf().setAppName("DStreamToMySQL")
        val ssc = new StreamingContext(conf,Seconds(1))
        //设置日志等级
        StreamingLoggingExample.setStreamingLogLevels()

        val lines = ssc.textFileStream("/tmp/yuhang.zhang/data")
        val words = lines.flatMap(_.split(" "))
        val pairWord = words.map((_,1))
        //累计更新
        val stateWordCount = pairWord.updateStateByKey[Int](updateFunc)

        //将stateWordCount存入数据库
        //stateWordCount中包含一堆的Rdd
        //我们需要对每个Rdd中的每条数据进行处理储存
        stateWordCount.foreachRDD(rdd => {
            //每个rdd中包含的数据类型为(String,Int)
            //我们把所有数据records定义为Iterator类型，方便我们遍历
            def func(records:Iterator[(String,Int)]): Unit ={
                //注意，conn和stmt定义为var不能是val
                var conn: Connection = null
                var stmt : PreparedStatement = null
                try{
                    //连接数据库
                    val url = "jdbc:mysql://localhost:3306/spark" //地址+数据库
                    val user = "root"
                    val password = ""
                    conn = DriverManager.getConnection(url,user,password)
                    //
                    records.foreach(p =>{
                        //wordcount为表名，word和count为要插入数据的属性
                        //插入数据
                        val sql = "insert into wordcount(word,count) values(?,?)"
                        stmt = conn.prepareStatement(sql)
                        stmt.setString(1,p._1.trim)
                        stmt.setInt(2,p._2.toInt)
                        stmt.executeUpdate()
                    })
                }catch {
                    case e : Exception => e.printStackTrace()
                }finally {
                    if(stmt != null)
                        stmt.close()
                    if(conn != null)
                        conn.close()
                }
            }
            val repairtitionedRDD = rdd.repartition(3)//将每个rdd重新分区
            repairtitionedRDD.foreachPartition(func)//对重新分区后的rdd执行func函数
        })
        ssc.start()//启动
        ssc.awaitTermination()//等待终止命令
    }

}

　　

转载于:https://www.cnblogs.com/zzhangyuhang/p/9075992.html