Avoiding loops in Scala

Question

I am very new to Scala and the whole functional programming style. What I need to do is calculating the similarity between two Strings by comparing each letter of both words. The function will be used with words of same length.

For example "network" and "workout" would have a similarity of 1. "House" and "Mouse" would have a similarity of 4.

Here is how I would do it in a very old fashioned C# way:

int calculateCharSimilarity(string first, string second)
{
  int similarity = 0;
  for(int i = 0; i < first.lenght() && i < first.lenght(); i++)
  {
    if(first.charAt(i) == second.charAt(i))
      similarity++;
  }
  return similarity;
}

What I did in scala so far is to write a tail recursive function in order to avoid the loop:

@tailrec
private def calculateCharSimilarity(first: Seq[Char], second: Seq[Char], similarity: Int = 0): Int = {
  if(first != Nil && second != Nil)
    calculateCharSimilarity(first.tail, second.tail, if(first.head == second.head) similarity + 1 else similarity)
  else
    similarity
}

But I am not so sure if that is best practise in Scala. Isn't there any way to do it more elegant with Collection Combinators (zip, filter) for example?

senia · Accepted Answer · 2014-01-06 19:21:04Z

10

def charSimilarity(first: String, second: String) =
  (first.view zip second).count{case (a, b) => a == b}

charSimilarity("network", "workout")
// Int = 1

charSimilarity("House", "Mouse")
// Int = 4

You could drop method view here. In this case you'll create a new collection of tuples (Char, Char) of size min(first.size, second.size). For small strings (single word) you'll get no performance issue.

Alternative implementation:

(first, second).zipped.count{case (a, b) => a == b}

edited Jan 6, 2014 at 19:21

answered Jan 6, 2014 at 18:57

senia

38.1k6 gold badges93 silver badges134 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

Patryk Ćwiek Over a year ago

For completeness and future reference, the body of the same method in C# could be return first.Zip(second, Tuple.Create).Count(t => t.Item1 == t.Item2);, so quite similar. :)

senia Over a year ago

@PatrykĆwiek: I guess first.Zip(second, (a, b) => a == b).Count(t => t) is even better (no unnecessary Tuple creations, and I just don't like ItemN and _N methods).

Patryk Ćwiek Over a year ago

True and good catch! Unfortunately, C# lacks the syntactic sugar and support for tuples akin to Scala and F# and I just wanted to be as close to your solution as possible.

Alexander Weber Over a year ago

Thanks @senia, that is exactly what I looked for! Just to follow up - what exactly is the benefit of 'view' in this case? Is it in order to create the tuples in a lazy way? Could an '(first.iterator zip second)' do it as well? PS: thanks PatrykĆwiek and senia, I didn't use the Zip extended method in C# so far, so I'll keep that one in mind!

senia Over a year ago

@AlexanderWeber: Could an '(first.iterator zip second)' do it as well?, yes, but you should use (first.iterator zip second.iterator). Iterator is mutable, view is immutable. It could lead to some strange errors, so I prefer to avoid iterators.

|

Collectives™ on Stack Overflow

Avoiding loops in Scala

1 Answer 1

7 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

7 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related