How to read CSV columns into vectors in Scala - scala

Question

I have a CSV file:
My CSV file
I want to create a map as such:
((A -> Vector(10.75,10.75,10.47,...,..), B-> Vector(164.56,164.99,160.98,...), C -> Vector(7.1,7.4,9.4,...,), D - > Vector(14.2,14.8,18.8,...,..))
This is what I have so far (No much):
val source=Source.fromFile("train3.csv")
var firstLine = source.getLines.find(_ => true).get
println(firstLine)
source.getLines().foreach(line=>{
val lines = line.split(",").map(_.trim).toVector
println(cols)
})
source.close()
This code prints this:
What my code prints
My Vectors contain the rows, I need my Vectors to contain the columns and not the rows from the CSV.

Answer 1

First off, please don't post important text information as an image. We can't copy-paste from images.
I created a smaller test file as follows:
A,B,C
10.75,10.75,10.47
164.56,164.99,160.98
7.1,7.4,9.4
14.2,14.8,18.8
With that in place I ran it through this code.
util.Using(io.Source.fromFile("train3.csv")){
_.getLines()
.map(_.split(","))
.toVector
.transpose
.groupMapReduce(_.head)(_.tail.map(_.toFloat))(_++_)
} //file is auto-closed
//res0: Try[Map[String,Vector[Float]]] =
// Success(Map(A -> Vector(10.75, 164.56, 7.1, 14.2)
// , B -> Vector(10.75, 164.99, 7.4, 14.8)
// , C -> Vector(10.47, 160.98, 9.4, 18.8)))
This will return Failure() if the file can't be opened, or the row lengths aren't consistent, or the number strings can't be converted. (Which might not be required. I just assumed it would be desirable.)

Answer 2

Once you have read the lines, and have something like
val lines: Seq[Seq[String]] = ???, just do
val columns = "ABC..."
lines
.iterator
.map { row => columns.zip(row) }
.groupBy { _._1 }
.mapValues { _._2 }
.toMap
If you want it specifically to be Vectors for some weird reason, then just add .to[Vector] after _._2