Github user DylanGuedes commented on a diff in the pull request:
https://github.com/apache/spark/pull/21045#discussion_r189421719
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/collectionOperations.scala
---
@@ -126,6 +126,149 @@ case class MapKeys(child: Expression)
override def prettyName: String = "map_keys"
}
+@ExpressionDescription(
+ usage = """_FUNC_(a1, a2, ...) - Returns a merged array containing in
the N-th position the
+ N-th value of each array given.""",
+ examples = """
+ Examples:
+ > SELECT _FUNC_(array(1, 2, 3), array(2, 3, 4));
+ [[1, 2], [2, 3], [3, 4]]
+ > SELECT _FUNC_(array(1, 2), array(2, 3), array(3, 4));
+ [[1, 2, 3], [2, 3, 4]]
+ """,
+ since = "2.4.0")
+case class Zip(children: Seq[Expression]) extends Expression with
ExpectsInputTypes {
+
+ override def inputTypes: Seq[AbstractDataType] =
Seq.fill(children.length)(ArrayType)
+
+ override def dataType: DataType = ArrayType(mountSchema)
+
+ override def prettyName: String = "zip"
+
+ override def nullable: Boolean = children.forall(_.nullable)
+
+ lazy val numberOfArrays: Int = children.length
+
+ private lazy val arrayTypes =
children.map(_.dataType.asInstanceOf[ArrayType])
+
+ private lazy val arrayElementTypes = arrayTypes.map(_.elementType)
+
+ def mountSchema: StructType = {
+ val fields = arrayTypes.zipWithIndex.foldRight(List[StructField]()) {
case ((arr, idx), list) =>
+ StructField(s"_$idx", arr.elementType, children(idx).nullable ||
arr.containsNull) :: list
+ }
+ StructType(fields)
+ }
+
+ override def doGenCode(ctx: CodegenContext, ev: ExprCode): ExprCode = {
+ val genericArrayData = classOf[GenericArrayData].getName
+ val genericInternalRow = classOf[GenericInternalRow].getName
+
+ val evals = children.map(_.genCode(ctx))
+
+ val arrVals = ctx.freshName("arrVals")
+ val arrCardinality = ctx.freshName("arrCardinality")
+ val biggestCardinality = ctx.freshName("biggestCardinality")
+ val storedArrTypes = ctx.freshName("storedArrTypes")
+
+ val inputs = evals.zipWithIndex.map { case (eval, index) =>
+ s"""
+ |${eval.code}
+ |if (!${eval.isNull}) {
+ | $arrVals[$index] = ${eval.value};
+ | $arrCardinality[$index] = ${eval.value}.numElements();
+ |} else {
+ | $arrVals[$index] = null;
+ | $arrCardinality[$index] = 0;
+ |}
+ |$storedArrTypes[$index] = "${arrayElementTypes(index)}";
+ |$biggestCardinality[0] = Math.max($biggestCardinality[0],
$arrCardinality[$index]);
+ """.stripMargin
+ }
+
+ val inputsSplitted = ctx.splitExpressions(
+ expressions = inputs,
+ funcName = "getInputAndCardinality",
+ arguments =
+ ("ArrayData[]", arrVals) ::
+ ("int[]", arrCardinality) ::
+ ("String[]", storedArrTypes) ::
+ ("int[]", biggestCardinality) :: Nil)
+
+ val myobject = ctx.freshName("myobject")
+ val j = ctx.freshName("j")
+ val i = ctx.freshName("i")
+ val args = ctx.freshName("args")
+
+ val fillValue = arrayElementTypes.distinct.map { case (elementType) =>
+ val getArrValsItem = CodeGenerator.getValue(s"$arrVals[$j]",
elementType, i)
+ s"""
+ | if ($storedArrTypes[$j] == "${elementType}") {
+ | $myobject[$j] = $getArrValsItem;
+ | }
+ """.stripMargin
+ }
+
+ val fillValueSplitted = ctx.splitExpressions(
+ expressions = fillValue,
+ funcName = "fillValue",
+ arguments =
+ ("int", j) ::
+ ("Array[] Object", myobject) :: Nil)
+
+ ev.copy(s"""
+ |ArrayData[] $arrVals = new ArrayData[$numberOfArrays];
+ |int[] $arrCardinality = new int[$numberOfArrays];
+ |int[] $biggestCardinality = new int[1];
--- End diff --
I used `int[]` because I thought that it was impossible to change
$biggestCardinality due to java not allowing passing int by reference. But I
found a workaround (inspired by `ConcatWs`).
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
