why can't we have capital letters in tuple variable declaration in scala

Question

Why can't we have Capital letter in a tuple variable, As I create a TestData it throws an error

val (trainingData, TestData): Tuple2[RDD[LabeledPoint],RDD[LabeledPoint]] = (splits(0), splits(1))

Answer 1

This is because you are not creating a ordinary variable in your case.

val (trainingData, TestData) = 

This is deconstructiong the right-hand side expression using the pattern matching on left-hand side. So the variables in pattern matching expressions must start with a lower-case letter.

val (trainingData, TestData) = (split(0), split(1)) 

will be equivalent to

(split(0), split(1)) match {
      case (trainingData, TestData) => (trainingData, TestData)
    }

Which tries to assign the two split values to trainingData, TestData which fails because it always accepts only with the lowercase variable at starting.

Rather, this will works fine

val (trainingData, testData) = (split(0), split(1))

This means,

(split(0), split(1)) match {
      case (trainingData, testData) => (trainingData, testData)
 }

Hope this helps!

Answer 2

In Scala you can create variables with a leading uppercase letter. (It's not recommended, but it can be done.)

val TestCnt = 7

But you can't do that when creating variables via pattern matching.

val (ch, num)       = ('x', 17)  // OK
val (Run, distance) = (true, 5)  // error: not found: value Run

Why is that? It's because the compiler needs to make a distinction between "constant patterns" and "variable patterns." This is explained in some detail in Section 15.2, "Kinds of Patterns" [PiS (1st Edition)], but the gist of it is that a leading uppercase letter is considered a constant, meaning that the pattern must match this value exactly, and a leading lowercase letter is considered a variable, which will match any value and that variable is also bound to the value.

someTuple match {
  case ('t', 42)    => /*will match only if both elements match these values*/
  case (_, TestCnt) => /*will match only if 2nd element same value as TestCnt*/
  case (c, n)       => /*will match any 2-ple, variables c,n bound to values*/
}

It's worth noting that there is a workaround for the lowercase-is-variable rule (use backticks) but there is no way to sidestep the uppercase-is-constant rule.