Scala的“=>”符号简介
Scala中的=>符号可以看做是创建函数实例的语法糖。例如:A => T,A,B => T表示一个函数的输入参数类型是“A”,“A,B”,返回值类型是T。请看下面这个实例:
scala> val f: Int => String = myInt => "The value of myInt is: " + myInt.toString()
f: Int => String =
scala> println(f(3))
The value of myInt is: 3
上面例子定义函数f:输入参数是整数类型,返回值是字符串。
另外,() => T表示函数输入参数为空,而A => Unit则表示函数没有返回值。
======================================================================
It?has several meanings in Scala, all related to its mathematical meaning as implication.
In a value, it introduces a?function literal(通译为匿名函数,有时候也叫函数显式声明,函数字面量), or?lambda(参考lambda表达式的文章,其实也是匿名函数).
e.g. the bit inside the curly braces in:?
List(1,2,3).map {?(x: Int) => x * 2?}
In a type, with symbols on both sides of the arrow (e.g.?A => T,?(A,B) => T,?(A,B,C) => T, etc.) it's sugar(syntactic sugar语法糖)?for?Function[A[,B,...],T], that is, a function that takes parameters of type?A[,B...], and returns a value of type?T.(语法糖通过更简洁的语法达到目的,直接把所需要的参数、类型、函数最简化,然后把解析的工作交给编译器来完成,这步称为去糖化。例如,(A,B)=>T,包含了函数,参数以及类型,实际上是一个匿名函数,func(A,B,T)或者func(A T,B T))。
=> is syntactic sugar for creating instances of functions. Recall that every function in scala is an instance of a class.
For example, the type Int => String, is equivalent to the type Function1[Int,String] i.e. a function that takes an argument of type Int and returns a String.?
scala>?val f: Function1[Int,String] = myInt => "my int: "+myInt.toString
? f: (Int) => String =
? scala> f(0)
? res0: String = my int: 0
? scala>?val f2: Int => String = myInt => "my int v2: "+myInt.toString
? f2: (Int) => String =
? scala> f2(1)
? res1: String = my int v2: 1
scala>?val f2: Function2[Int,Int,String] = (myInt1,myInt2) => "This is my function to transfer " + myInt1 + " and " + myInt2 + " as a string component."
f2: (Int, Int) => String =
scala> f2(1,2)
res6: String = This is my function to transfer 1 and 2 as a string component.
scala>?val f22:(Int,Int)=>String = (myInt1,myInt2) => "This is my function to transfer " + myInt1 + " and " + myInt2 + " as a string component."
f22: (Int, Int) => String =
scala> f22(2,4)
res7: String = This is my function to transfer 2 and 4 as a string component.
Here myInt is binded to the argument value passed to f and f2.
() => T?is the type of a function that takes no arguments and returns a?T. It is equivalent to?Function0[T].?()?is called a zero parameter list I believe.
scala>?val f:?() => Unit?=?() => { println("x")}
f: () => Unit =
scala> f()
x
scala> val f2: Function0[Unit] = () => println("x2")
f: () => Unit =
scala> f2()
x2
As the most simplified answer, you can substitute whatever is on the left-hand side of => with the word "LEFT" and whatever is on the right-hand side with the word "RIGHT".
Then, the meaning of "LEFT => RIGHT" becomes:
Take LEFT then do RIGHT.
This means that if you have a "()=>" that you can take nothing (that is, no parameters) and then do whatever is on the right-hand side.
This is the most common meaning.
Empty parens on the left hand side?(e.g.?() => T) indicate that the function takes no parameters (also sometimes called a "thunk");
? ? ? ?
定时器的示例程序还可以做一些改进。首先,timeFlies 函数只被用过一次,也就是当做回调函数传给 oncePerSecond 的时候,对于这种函数,在用到的时候即时构造更合理,因为可以省去定义和命名的麻烦,在 Scala 中,这样的函数称为匿名函数(anonymousfunctions),也就是没有名字的函数。使用匿名函数代替 timeFlies 函数后的程序代码如下:
?object TimerAnonymous {
def oncePerSecond(callback: () => Unit) {
while (true) { callback(); Thread sleep 1000 }
?}
?
def main(args: Array[String]) {
oncePerSecond(() =>
println("time flies like an arrow..."))
?}
}
代码中的右箭头‘=>’表明程序中存在一个匿名函数,箭头左边是匿名函数的参数列表,右边是函数体。在本例中,参数列表为空(箭头左边是一对空括号),而函数体和改造前定义的 timeFlies 函数保持一致。
函数求值有两种策略,一种是传值调用,一种是传名调用,其中编译器的"传名调用"实现,往往是将参数放到一个临时函数之中,再将这个临时函数传入函数体。这个临时函数就叫做 Thunk 函数,也称为柯里化。
function f(m){
? return m * 2; ??
}
f(x + 5);
// 等同于
var thunk = function () {
? return x + 5;
};
function f(thunk){
? return thunk() * 2;
}
Empty parens on the right hand side?denote that it returns?()?— the sole value of type?Unit, whose name can also be written?()—confused yet? :)
A function that returns Unit is also known as a?procedure, normally a method that's called only for its?side effect.
Scala对于不返回值的函数有特殊的表示法。如果函数体包含在花括号当中但没有前面的=号,那么返回值类型就是Unit。这样的函数被称做过程(procedure)。过程不返回值,我们调用它仅仅是为了它的副作用。例如:
scala> def box(s: String) {
? ? | var border = "-" * s.length + "--\n"
? ? | println(border + "|" + s + "|\n" + border)
? ? | }
box: (s: String)Unit
scala> box("Hello")
-------
|Hello|
-------
关于side effect的翻译
点击了一个按钮,用户所期待的行为不叫“side effects”,用户没有期待的行为才叫“side effects”。
函数一般认为是没有状态的,每次用相同的输入调用时应当给出相同的输出,且不会影响程序的其它部分。如果函数对程序的状态有影响,这种影响就叫作“side effects”。
总之,“side effects”指的就是那些本不应有或者用户意料之外的作用,翻译成“副作用”我觉得挺恰当的。它可以看成医学上“副作用”一词的合理引申。
举个简单的例子好了:
int counter = 0;
// with side effectint incCounter() {
? ? counter += 1;
? ? return counter;}
// without side effectint incNumber(int m) {
? ? return m + 1;}
incCounter 修改了它 scope 以外的东西,它跑过一遍之后会修改外部环境,我们讲这叫 side effect。intNumber 跑完之后本来什么样还是什么样,这就叫没有 side effects。翻成「副作用」没有问题。
In the type declaration for a method or function parameter, with no symbol on the left hand side?(e.g.?def f(param: => T)) it's a "by-name parameter"(传名参数,或者翻译为叫名参数), meaning that is evaluated every time it's used within the body of the function, and not before. Ordinary?"by-value" parameters?(传值参数)are evaluated before entry into the function/method.
By-Value参数,我们都比较熟悉,参数在进入函数前就进行计算,最后传入的最终的计算结果。
By-Name参数,表示参数在进入函数后,每次在函数体内调用的时候才会计算。
def myByName1(param:?() => Boolean): Unit = {
? ? println("start")
? ? println(param)? ? ? //println(param.apply())也可以
? ? println("end")
? }
//要实现传名函数,要定义的类型开始于“=>”,而不是“() =>”。
//myByName1{7>9}不会有效,因为缺少“() =>”
myByName1{() =>
? ? println("enter")
? ? 7 > 9
}
In a?case?clause(case子句), they separate the pattern (and optional guard) from the result expression, e.g.?case x => y.
switch
val friend =
? firstArg match {
? case "salt" => "pepper"
? case "chips" => "salsa"
? case "eggs" => "bacon"
??case ?_ => "huh?"
? }
exception
try {
? val file = new FileReader("input.txt")
} catch {
??case ex: FileNotFoundException =>?ex.printStackTrace()
? case ex: IOException =>?? //ex.printStackTrace()可以省略
} finally {
? file.close()
}
这个try-catch表达式的行为与其它语言中的异常处理一致。程序体被执行,如果抛出异常,每个catch子句依次被尝试。本例中,如果异常是FileNotFoundException,那么第一个子句将被执行。如果是IOException类型,第二个子句将被执行。如果都不是,那么try-catch将终结并把异常上升出去。
注意
你将很快发现与Java的一个差别是Scala里不需要你捕获检查异常:checked exception,或把它们声明在throws子句中。如果你愿意,可以用ATthrows标注声明一个throws子句,但这不是必需的。