Github user DylanGuedes commented on a diff in the pull request:
https://github.com/apache/spark/pull/21045#discussion_r190014678
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/collectionOperations.scala
---
@@ -127,6 +127,148 @@ 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 = Math.max($biggestCardinality,
$arrCardinality[$index]);
+ """.stripMargin
+ }
+
+ val inputsSplitted = ctx.splitExpressions(
+ expressions = inputs,
+ funcName = "getInputAndCardinality",
+ returnType = "int",
+ makeSplitFunction = body =>
+ s"""
+ |$body
+ |return $biggestCardinality;
+ """.stripMargin,
+ foldFunctions = _.map(funcCall => s"$biggestCardinality =
$funcCall;").mkString("\n"),
+ arguments =
+ ("ArrayData[]", arrVals) ::
+ ("int[]", arrCardinality) ::
+ ("String[]", storedArrTypes) ::
+ ("int", biggestCardinality) :: Nil)
+
+ val myobject = ctx.freshName("myobject")
+ val j = ctx.freshName("j")
+ val i = ctx.freshName("i")
+ val args = ctx.freshName("args")
+
+ val fillValue = arrayElementTypes.distinct.map { case (elementType) =>
+ val getArrValsItem = CodeGenerator.getValue(s"$arrVals[$j]",
elementType, i)
+ s"""
+ |if ($storedArrTypes[$j] == "${elementType}") {
+ | $myobject[$j] = $getArrValsItem;
+ |}
+ """.stripMargin
+ }.mkString("\n")
+
+ ev.copy(s"""
+ |ArrayData[] $arrVals = new ArrayData[$numberOfArrays];
+ |int[] $arrCardinality = new int[$numberOfArrays];
+ |int $biggestCardinality = 0;
+ |String[] $storedArrTypes = new String[$numberOfArrays];
+ |$inputsSplitted
+ |Object[] $args = new Object[$biggestCardinality];
+ |for (int $i = 0; $i < $biggestCardinality; $i ++) {
+ | Object[] $myobject = new Object[$numberOfArrays];
+ | for (int $j = 0; $j < $numberOfArrays; $j ++) {
+ | if ($arrVals[$j] != null && $arrCardinality[$j] > $i &&
!$arrVals[$j].isNullAt($i)) {
+ | $fillValue
+ | } else {
+ | $myobject[$j] = null;
+ | }
+ | }
+ | $args[$i] = new $genericInternalRow($myobject);
+ |}
+ |boolean ${ev.isNull} = false;
--- End diff --
I didn't checked how Presto handle this case, good to know. I'll change
them to also return null in such cases.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
