00.scala2014. 3. 28. 13:40
반응형

The Scala List class filter method implicitly loops over the List you supply, tests each element of the List with the function you supply. Your function must return true or false, and filter returns the list elements where your function returns true.

(Note: Even though I use a List in these examples, the filter method can be used on any Scala sequence, including Array, List, Vector, Seq, etc.)

Let's look at a few simple examples. In this first example we filter a small list of numbers so that our resulting list only has numbers that are greater than 2:

scala> val nums = List(5, 1, 4, 3, 2)
nums: List[Int] = List(5, 1, 4, 3, 2)

scala> nums.filter(_ > 2)
res0: List[Int] = List(5, 4, 3)

Note that in the real world you'd assign the filtered results to a new List, like this:

val originalList = List(5, 1, 4, 3, 2)
val newList = originalList.filter(_ > 2)

This example shows how to get the even numbers from a List using a simple modulus test:

scala> nums.filter( _ % 2 == 0 )
res21: List[Int] = List(4, 2)

You can take that example a step further by filtering and then sorting the list:

# filter and sort
scala> nums.filter( _ % 2 == 0 ).sort(_ < _)
warning: there were 1 deprecation warnings; re-run with -deprecation for details
res22: List[Int] = List(2, 4)

filter method examples with a List of Strings

Here are two filter method examples with a list of Strings:

val fruits = List("orange", "peach", "apple", "banana")

scala> fruits.filter(_.length > 5)
res21: List[java.lang.String] = List(banana, orange)

scala> fruits.filter(_.startsWith("a"))
res22: List[java.lang.String] = List(apple)

Combining filter, sort, and map

From the excellent book, Beginning Scala, here's a nice combination of the List filter, sort, and map methods:

trait Person {
  def first: String
  def age: Int
  def valid: Boolean
}

Returns the first name of 'valid' persons, sorted by age

def validByAge(in: List[Person]) =
  in.filter(_.valid).
  sort(_.age < _.age).
  map(_.first)

The following example shows how you can use filter with map to transform the type of data that the expression returns. In this case we'll start with a sequence of Person objects, and transform it into a sequence of String objects.

We'll start with a simple case class:

scala> case class Person(first: String, last: String, mi: String)
defined class Person

Next, we'll create a little sequence of Person objects:

scala> val fred = Person("Fred", "Flintstone", "J")
fred: Person = Person(Fred,Flintstone,J)

scala> val wilma = Person("Wilma", "Flintstone", "A")
wilma: Person = Person(Wilma,Flintstone,A)

scala> val barney = Person("Barney", "Rubble", "J")
barney: Person = Person(Barney,Rubble,J)

scala> val betty = Person("Betty", "Rubble", "A")
betty: Person = Person(Betty,Rubble,A)

scala> val peeps = Seq(fred, wilma, barney, betty)
peeps: Seq[Person] = List(Person(Fred,Flintstone,J), Person(Wilma,Flintstone,A), Person(Barney,Rubble,J), Person(Betty,Rubble,A))

Finally, we'll combine filter and map to get a list of all first names where the last name is "Flintstone":

scala> peeps.filter(_.last == "Flintstone").map(_.first)
res0: Seq[String] = List(Fred, Wilma)

The way this works is:

  • The filter method returns a sequence of Person objects where the last name is "Flintstone".
  • The map method call gets the first name of each Person object. This results in a sequence of strings, where each string is the first name of each person that came out of the filter call.

I initially wrote this as a for/yield loop, but then realized I could write this much more concisely with this approach. At the moment I find the for/yield loop to be more readable, and this to be much more concise.

In my opinion, this code can be made a little more readable by using a variable name in the mapexpression, as a reminder that you're still dealing with Person objects:

scala> peeps.filter(_.last == "Flintstone").map(person => person.first)
res1: Seq[String] = List(Fred, Wilma)

Scala List filter method summary

I hope these filter method examples have been helpful. Here's a quick summary of how the filter method works:

  • filter implicitly loops over a List.
  • filter takes a function as an argument. That function should take one List element as input, perform the test you define, and then return either true or false (a Boolean).
  • filter only returns List elements that match the filtering expression.


Posted by 1010