Spark和Spring整合处理离线数据

object Bootstrap {
  private val log = LoggerFactory.getLogger(Bootstrap.getClass)

  //指定配置文件如log4j的路径
  val ConfFileName = "conf"
  val ConfigurePath = new File("").getAbsolutePath.substring(0, if (new File("").getAbsolutePath.lastIndexOf("lib") == -1) 0
  else new File("").getAbsolutePath.lastIndexOf("lib")) + this.ConfFileName + File.separator

  //存放实现了StatsTask的离线程序处理的类
  private val TASK_MAP = Map("WordCount" -> classOf[WordCount])

  def main(args: Array[String]): Unit = {
    //传入一些参数，比如要运行的离线处理程序类名、处理哪些时间的数据
    if (args.length < 1) {
      log.warn("args 参数异常！！！" + args.toBuffer)
      System.exit(1)
    }
    init(args)
  }

  def init(args: Array[String]) {
    try {
      SpringUtils.init(Array[String]("applicationContext.xml"))
      initLog4j()

      val className = args(0)
      // 实例化离线处理类
      val task = SpringUtils.getBean(TASK_MAP(className))

      args.length match {
        case 3 =>
          // 处理一段时间的每天离线数据
          val dtStart = DateTimeFormat.forPattern("yyyy-MM-dd").parseDateTime(args(1))
          val dtEnd = DateTimeFormat.forPattern("yyyy-MM-dd").parseDateTime(args(2))
          val days = Days.daysBetween(dtStart, dtEnd).getDays + 1
          for (i <- 0 until days) {
            val etime = dtStart.plusDays(i).toString("yyyy-MM-dd")
            task.runTask(etime)

            log.info(s"JOB --> $className 已成功处理: $etime 的数据")
          }

        case 2 =>
          // 处理指定的某天离线数据
          val etime = DateTimeFormat.forPattern("yyyy-MM-dd").parseDateTime(args(1)).toString("yyyy-MM-dd")
          task.runTask(etime)
          log.info(s"JOB --> $className 已成功处理: $etime 的数据")

        case 1 =>
          // 处理前一天离线数据
          val etime = DateTime.now().minusDays(1).toString("yyyy-MM-dd")
          task.runTask(etime)
          log.info(s"JOB --> $className 已成功处理: $etime 的数据")

        case _ => println("执行失败 args参数:" + args.toBuffer)
      }
    } catch {
      case e: Exception =>
        println("执行失败 args参数:" + args.toBuffer)
        e.printStackTrace()
    }

    // 初始化log4j
    def initLog4j() {
      val fileName = ConfigurePath + "log4j.properties"
      if (new File(fileName).exists) {
        PropertyConfigurator.configure(fileName)
        log.info("日志log4j已经启动")
      }
    }
  }
}

object SpringUtils {
  private var context: ClassPathXmlApplicationContext = _

  def getBean(name: String): Any = context.getBean(name)

  def getBean[T](name: String, classObj: Class[T]): T = context.getBean(name, classObj)

  def getBean[T](_class: Class[T]): T = context.getBean(_class)

  def init(springXml: Array[String]): Unit = {
    if (springXml == null || springXml.isEmpty) {
      try
        throw new Exception("springXml 不可为空")
      catch {
        case e: Exception => e.printStackTrace()
      }
    }
    context = new ClassPathXmlApplicationContext(springXml(0))
    context.start()
  }

}

<?xml version="1.0" encoding="UTF-8"?>
<beans xmlns="http://www.springframework.org/schema/beans"
       xmlns:context="http://www.springframework.org/schema/context"
       xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
       xsi:schemaLocation="http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans-4.0.xsd
          http://www.springframework.org/schema/context  http://www.springframework.org/schema/context/spring-context-4.0.xsd">

    <!-- 配置包扫描 -->
    <context:component-scan base-package="com.bigdata.stats"/>

</beans>

trait StatsTask extends Serializable {
  //"子类"继承StatsTask重写该方法实现自己的业务处理逻辑 
  def runTask(etime: String)
}

//不要忘记添加 @Component ，否则无法利用Spring对WordCount进行实例化
@Component
class WordCount extends StatsTask {

  override def runTask(etime: String): Unit = {
    val sparkSession = SparkSession
      .builder()
      .appName("test")
      .master("local[*]")
      .getOrCreate()

    import sparkSession.implicits._

    val words = sparkSession.read.textFile("/Users/BigData/Documents/data/wordcount.txt").flatMap(_.split(" "))
      .toDF("word")

    words.createOrReplaceTempView("wordcount")

    val df = sparkSession.sql("select word, count(*) count from wordcount group by word")

    df.show()
  }
}