quick comparison in processing collections in Ruby vs Scala

collection.rb

start = Time.now
100.times do
  letter_list = []
  alphabet = ('a'..'z').to_a
  100_000.times{ letter_list << alphabet[rand(26)] }
  pick = letter_list.uniq[rand(26)]
  before, after = letter_list.partition{|n| n > pick}
end
p (Time.now - start) * 1000

collection.scala

import scala.util.Random

val start = System.currentTimeMillis
(1 to 100).foreach { _ =>
  val alphabet = ('a' to 'z').toArray
  val letterList = (0 to 100000).map( _ => alphabet(Random.nextInt(26)) )
  val pick = letterList.distinct(Random.nextInt(26))
  val (before, after) = letterList.partition{ _ > pick }
}
println(System.currentTimeMillis - start)

Not your usual scala code but I was just curious how collection performed in scala vs jruby.

Scala: 787ms
jRuby 1.7 (JDK 7): 3328.0ms

You can shave off ~15% by using Array#sample instead of rand(26) :-)

I know this is really old, but it's a top Google result so thought I'd share. Using a solid benchmarking tool like Caliper to allow for JVM warmup will produce results about 200X faster for the Scala version: https://github.com/sam/collections-vs (<4ms instead of ~800ms).