This is blog entry originally written in the middle May 2014. I also seem to writing this code every 3 months or so and then I forgot. So I need a trigger to remember exactly how to work Scala Maps and sort the entries by key.

Let’s break out the Scala REPL.

Welcome to Scala version 2.10.4 (Java HotSpot(TM) 64-Bit Server VM, Java 1.7.0_51).
Type in expressions to have them evaluated.
Type :help for more information.

We will create a map of collection of String associated with Integer elements. This data might have come a key value storage or you may have a similar requirement to work with statistical data.

scala> val map = Map( "Jane" -> 3, "Peter" -> 10, "Steve" -> 1, "Anna" -> 5, "Megan" -> 15, "Brian" -> 7, "Sally" ->  8 )
map: scala.collection.immutable.Map[String,Int] = Map(Megan -> 15, Anna -> 5, Jane -> 3, Brian -> 7, Steve -> 1, Sally -> 8, Peter -> 10)

In order to sort the data structure, we then convert the map to list collection of tuples.

scala> map.toList
res1: List[(String, Int)] = List((Megan,15), (Anna,5), (Jane,3), (Brian,7), (Steve,1), (Sally,8), (Peter,10))

Now we can sort using the tuples. Here is the descending order on the values:

scala> map.toList.sortWith( (x,y) => x._2 > y._2 )
res2: List[(String, Int)] = List((Megan,15), (Peter,10), (Sally,8), (Brian,7), (Anna,5), (Jane,3), (Steve,1))

Here is the ascending order on the values

scala> map.toList.sortWith( (x,y) => x._2 < y._2 )
res3: List[(String, Int)] = List((Steve,1), (Jane,3), (Anna,5), (Brian,7), (Sally,8), (Peter,10), (Megan,15))

Here is another tip. Let’s say you need to create a histogram of 100 different product items. You can create this data structure
immutable at the beginning. However, as you are building the statistic, you may prefer to use an mutable collection map instead.

 val buckets = collection.mutable.Map.empty[Int,Int]
    buckets ++= (1 to 100).toList.map{ x => (x,0) }.toMap

Now in your statistic gathering part, you can write something like this:

val keyName: String = ???
    val keyIndex: Integer = convertNameToIndex(keyName)
    buckets(keyIndex) =  buckets.getOrElse(keyIndex, 0) + 1

And then render the sorted data into a top ten products

val sortedProducts = buckets.toList.sortWith{ (x,y) => x._2 > y._2 }
 println(sortedProducts.take(10))

That’s all.