1.模式匹配比java中的switch case强大很多,除了值,类型,集合等进行匹配,最常见的Case class进行匹配,Master.scala有大量的模式匹配。
Case "_"表示不满足上面的所有情况的体验,举个例子:
def bigdata(data: String){
data match{
case "Spack"=>println("WOW!!!")
case "Hadoop"=>println("OK")
case _=>println("sorry")
}
}
bigdata("Spack") //Wow!!!
可以在case里面加入条件判断
def bigData(data:String)
{
data match{
case "Spark"=>println("Wow!!")
case "hadoop"=> println("ok")
case _ if data == "Flink" =>println("Flink")
case _=>println("other")
}
}
bigdata("Flink") //Flink
对类型进行匹配
def exception(e: Exception){
e match{
casefileException:FileNotFoundException =>println("File not fount : " +fileException)
case _:Exception=>println("Exception" ,e) }
}
exception(new FileNotFoundException("oop!!!")) //> File notfount : java.io.FileNotFoundException: oop!!!
对集合进行匹配
def data (array:Array[String])
{ arraymatch{
case Array("Scala")=>println("Scala")
caseArray(spark,hadoop,flink)=>println(spark +" : " +hadoop +" : " +flink +" : ")
case Array("Spark",_*)=>println("Spark...")
case _=>println("Unkown")
}
} //> data: (array: Array[String])Unit
data(Array("Spark")) //> Spark...
data(Array("Scala")) //> Scala
data(Array("Scala","Spark","kafaka")) //> Scala : Spark : kafaka :
对class进行匹配
scala> case class Person(name: String)
defined class Person
case classPerson(name: String)
Person("Spark") // res0: worksheetest.Person = Person(Spark)
1:case class 相对于java中的bean,val 只有个get
2:实例自动调用伴生对象
class Person
case classWorker(name: String,salary: Double) extends Person
case classStudent(name: String,score: Double) extends Person
def sayHi(person:Person)
{
personmatch{
case Student(name,score)=>println("I am Student :"+name +","+score)
case Worker(name,salary)=>println("I am Worker :"+name +","+salary)
case _ =>println("Unknown")
}
} //> sayHi: (person: worksheetest.Person)Unit
sayHi(Worker("Worker",6.5)) //> I am Worker :Worker,6.5
sayHi(Student("Student",6.5)) //> I am Student :Student,6.5
DeployMessages源码中:
caseclassExecutorStateChanged(
appId:String,
execId:Int,
state:ExecutorState,
message:Option[String],
exitStatus:Option[Int])
extends DeployMessage
case class 使用时会生成很多对象
case object 本身就是一个实例,全局唯一
scala 的类型参数(重磅的东西)最好的难点,太有用了,在所有的spark源码中到处都是
例:RDD[T: ClassTag]
泛型,参数本身是有类型,scala的泛型,
泛型类和泛型函数
class Person[T](valcontent:T)
{
def getContent(id: T) = id+ " _ "+ content
}
val p = newPerson[String]("Spark") //> p :worksheetest.Person[String] = worksheetest$Person@50134894
p.getContent("Scala") //> res0: String = Scala _ Spark
泛型前面有+和-
* scala> def mkArray[T : ClassTag](elems: T*) =Array[T](elems: _*)
* mkArray: [T](elems: T*)(implicit evidence$1:scala.reflect.ClassTag[T])Array[T]
*
* scala> mkArray(42, 13)
* res0: Array[Int] = Array(42, 13)
*
* scala> mkArray("Japan","Brazil","Germany")
* res1: Array[String] = Array(Japan, Brazil, Germany)
* }}}
协变:如果S是T的子类型,并且List[S]也是List[T]的子类型,那么成为协变 class Person[+T] //强制定义为协变类型
C[+T]:如果A是B的子类,那么C[A]是C[B]的子类。逆变范围小
C[-T]:如果A是B的子类,那么C[B]是C[A]的子类。协变 范围大
C[T]:无论A和B是什么关系,C[A]和C[B]没有从属关系。
注:阅读Spark源码 RDD、HadoopRDD、SparkContext、Master、Worker的源码,并分析里面使用的所有的模式匹配和类型参数的内容。
总结:
T <% Writable: ClassTag
T可以隐身转换为Writable类型
ClassTag在上下文中注入隐式值
对于Manifest Context Bounds
[T : Manifest] 进化为ClassTag了,T:ClassTag 运行时传递完整的类型上下文信息
Seq[Dependency[_]] 相当于Seq[Dependency[T]]
另外有段重要注释:
{{{
* scala> def mkArray[T : ClassTag](elems: T*) = Array[T](elems: _*)
* mkArray: [T](elems: T*)(implicit evidence$1: scala.reflect.ClassTag[T])Array[T]
*
* scala> mkArray(42, 13)
* res0: Array[Int] = Array(42, 13)
*
* scala> mkArray("Japan","Brazil","Germany")
* res1: Array[String] = Array(Japan, Brazil, Germany)
* }}}
*
表明了ClassTag 的隐式转换机制。