Map letter to word in which it appears - scala

Question

I am newbie to scala and I am trying to write a function that takes an input string and returns a map of letters to words in which they appear.
For example given the input string "this is demo",
I would like the output map ['t'->["this"],'h'->["this],'i'->["this","is"]... and so on.
I can write this code in traditional way, but how can I write this code in a functional way by using scala constructs like map, groupby, flatmap etc?

Answer 1

"this is demo"
.split(" ")
.flatMap(w => w.map(c => c -> w))
.groupMap(_._1)(_._2)
// HashMap(e -> Array(demo), s -> Array(this, is), t -> Array(this), m -> Array(demo), i -> Array(this, is), h -> Array(this), o -> Array(demo), d -> Array(demo))
First step consists in getting an array of tuples representing for each character which word it comes from. This can be achieved by first splitting the sentence in words and for each character of each word producing a tuple with the character and its word (.map(w => w.map(c => c -> w))). And since this gives us an array of arrays, we can use a flatMap to flatten these into a one level array of tuples (producing Array((t,this), (h,this), (i,this), ...)).
Second step consists in grouping these tuples of characters and words by character and mapping the grouped values to the associated words. Which can be achieved with groupMap (it groups tuples by their first part (by character) and maps grouped tuples to their second part (the word)). If you're using an earlier Scala version (before 2.13), you'll have to replace groupMap with a combination of groupBy and mapValues: .groupBy(_._1).mapValues(_.map(_._2)).

Answer 2

Here is another solution. You can replace the rough tokenizer I suggest by a parser, like Stanford CoreNLP Simple for example, info here.
def tokenizeText(s:String):Array[String] = {
s.toLowerCase().split("[\\W]+")
}
val text:String = "Here you have your text. Set several sentences if you like."
val words = tokenizeText(text)
val letters = words.mkString("").toSet.mkString("").split("")
val twoDArray:Array[Array[String]] = letters.map(l => words.filter(w => w.contains(l)))