Scala 元组添加保持序列相同的顺序

Question

I want to add scala tuples in respective order, it should add to list in its sequence

val d = (List.empty[String],List.empty[String],List.empty[String])
("1","2","3") :: d 
("4","5","6") :: d 

d should give output like (List("1","4"), List("2","5"),List("3","6"))

I have header data of below type and from the entity, I want to extract data of few attributes and accumulate as a tuple, I tried below approach which extracts data from the entity but not matching missing data type

header type Task[Map[String,List[Entity]]]

val missing:(Set[String],Set[String],Set[String]) = headers.map(_.foldLeft((Set.empty[String],Set.empty[String],Set.empty[String))((a,v)=> {
          v._2.map {
            entity =>{
              val x = entity.field1++a._1
              val y = entity.field2++a._2
              val z = entity.field3++a._3
              (x,y,z)
            }
           }
        }
   ))

Above data type does not match and get compilation error.

Answer 1

Something like this?

implicit class TriplePrepend[A](t :(List[A],List[A],List[A])) {
  def :: (x :(A,A,A)) = (x._1 :: t._1, x._2 :: t._2, x._3 :: t._3)
}

val d = (List.empty[String],List.empty[String],List.empty[String])

("1","2","3") :: ("4","5","6") :: d 
//res0: (List(1, 4),List(2, 5),List(3, 6))

Answer 2

You could use shapeless, which gives you the possibility to create generic functions working over tuples with various arity. First, you need to create functions:

import shapeless.syntax.std.tuple._
import cats.implicits._

object lifter extends (Id ~> List) { //shapeless polymorphic function lifting value into List
  override def apply[T](f: Id[T]): List[T] = List(f)
}

object folder extends Poly1 { //function which folds two lists using monoid
  implicit def caseTupleOfLists[A](implicit monoid: Monoid[A]) = at[(List[A], List[A])]{case (a, b) => a |+| b}
}

Then you can use it like this:

val a = ("1", "2", "3").map(lifter) //(List(1),List(2),List(3))
val b = ("4", "5", "6").map(lifter) //(List(4),List(5),List(6))

val result = a.zip(b).map(folder) //(List(1, 4),List(2, 5),List(3, 6))

One benefit over jwvh's solutions is, that it will work also on tuples with different arity:

val a2 = ("1", "2", "3", "foo").map(lifter)
val b2 = ("4", "5", "6", "bar").map(lifter)

val result2 = a2.zip(b2).map(folder) //(List(1, 4),List(2, 5),List(3, 6),List(foo, bar))

Update

If you map over Task (whether it's from ZIO, Monix or another library) you can't just get a result like (Set[String],Set[String],Set[String]) . You can only get Task[(Set[String], Set[String], Set[String])] . I modified your code to use pattern matching (it's usually more readable this way) and now it compiles:

case class Entity(field1: String, field2: String, field3: String)

val headers: Task[Map[String, List[Entity]]] = ???

val missing: Task[(Set[String], Set[String], Set[String])] =
  headers.map(_.foldLeft((Set.empty[String], Set.empty[String], Set.empty[String])) {
    case ((fields1, fields2, fields3), _ -> Entity(field1, field2, field3)) => {
      val x = fields1 + field1
      val y = fields2 + field2
      val z = fields3 + field3
      (x, y, z)
  }
})