Defective schema generation on array/seq column

[tribbloid]

To reproduce:

object NestedFieldWithColumn {

  case class Inner(
      v: Int,
      a: String,
      b: String
  )

  case class Outer1(
      in: Inner,
      aux1: String,
      aux2: String
  )

  case class OuterN(
      ins: Seq[Inner],
      aux1: String,
      aux2: String
  )

}

      val rows = Seq(
        OuterN(
          Seq(
            Inner(1, "a", "b"),
            Inner(1, "a", "b")
          ),
          "c",
          "d"
        ),
        OuterN(
          Seq(
            Inner(1, "a", "b"),
            Inner(1, "a", "b")
          ),
          "c",
          "d"
        )
      )

      val ds2 = TypedDataset.create(rows)

this will cause the error java.lang.NoSuchMethodError: 'boolean org.apache.spark.sql.catalyst.ScalaReflection$.isNativeType(org.apache.spark.sql.types.DataType)'

Looks like a version compatibility issue but it is not, the same error will be thrown for any Spark v3.4.0+

[tribbloid]

Ah, looks like its a duplicate of #755, these 2 should be closed simultaneously

[pomadchin]

Hey @tribbloid, what Spark and Frameless version do you use?

[tribbloid]

3.5.0 & 0.15.0 😱

[pomadchin]

Haha sounds about right, we need to cut a 3.5.x release.

So sad that they broke API again.

[pomadchin]

Closed via #772