typelevel#803 - base good case

dataset/src/test/scala/frameless/functions/UdfTests.scala

@@ -1,6 +1,7 @@
 package frameless
 package functions
 
+import org.apache.spark.sql.catalyst.analysis.UnresolvedAttribute
 import org.scalacheck.Prop
 import org.scalacheck.Prop._
 
@@ -62,8 +63,8 @@ class UdfTests extends TypedDatasetSuite {
       (dataset21 ?= d) && (dataset22 ?= d)
     }
 
-    check(forAll(prop[Int, Int, Int] _))
-    check(forAll(prop[String, Int, Int] _))
+    //check(forAll(prop[Int, Int, Int] _))
+    //check(forAll(prop[String, Int, Int] _))
     check(forAll(prop[X3[Int, String, Boolean], Int, Int] _))
     check(forAll(prop[X3U[Int, String, Boolean], Int, Int] _))
   }
@@ -182,4 +183,28 @@ class UdfTests extends TypedDatasetSuite {
 
     check(forAll(prop[Int, Int, Int, Int, Int] _))
   }
+
+  test("struct of arrays explode and star should resolve with udf") {
+
+    val dataset = TypedDataset.create(Seq(
+      X2(1, Seq(X2(1, "a"),X2(2, "b"),X2(5, "c"),X2(6, "d"))),
+      X2(2, Seq(X2(7, "a"),X2(8, "b"),X2(9, "c"),X2(10, "d"))),
+      X2(3, Seq(X2(11, "a"),X2(12, "b"),X2(13, "c"),X2(14, "d")))
+    ))
+    import org.apache.spark.sql.functions.{expr, explode => sexplode}
+    import org.apache.spark.sql.Column
+
+    val starter = UnresolvedAttribute("struct")
+
+    val inc = FramelessUdf[X2[Int, String], X2[Int, String]](
+      (x: X2[Int, String]) => x.copy(x.a + 1),
+      encoders = List(TypedEncoder[X2[Int, String]]),
+      Seq(TypedEncoder[X2[Int, String]].fromCatalyst(starter)), // b doesn't resolve
+      TypedEncoder[X2[Int, String]])
+
+//    dataset.dataset.select(expr("*"), sexplode(new Column(inc)).as("struct")).select(expr("struct.*")).show
+    dataset.dataset.select(expr("*"), sexplode(new Column("b")).as("struct")).
+       select(new Column(inc).as("struct")).
+      select(expr("struct.*")).show
+  }
 }