[GitHub] [spark] zhengruifeng commented on a diff in pull request #40563: [SPARK-41232][SPARK-41233][FOLLOWUP] Refactor `array_append` and `array_prepend` with `RuntimeReplaceable`

["zhengruifeng (via GitHub)" <gi...@apache.org>]

zhengruifeng commented on code in PR #40563:
URL: https://github.com/apache/spark/pull/40563#discussion_r1155685702


##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/collectionOperations.scala:
##########
@@ -5056,128 +4950,45 @@ case class ArrayCompact(child: Expression)
   """,
   since = "3.4.0",
   group = "array_funcs")
-case class ArrayAppend(left: Expression, right: Expression)
-  extends BinaryExpression
-    with ImplicitCastInputTypes
-    with ComplexTypeMergingExpression
-    with QueryErrorsBase {
-  override def prettyName: String = "array_append"
+case class ArrayAppend(left: Expression, right: Expression) extends InsertArrayOneSide {
 
-  @transient protected lazy val elementType: DataType =
-    inputTypes.head.asInstanceOf[ArrayType].elementType
+  override lazy val replacement: Expression =
+    ArrayInsert(left, Add(ArraySize(left), Literal(1)), right)
 
-  override def inputTypes: Seq[AbstractDataType] = {
-    (left.dataType, right.dataType) match {
-      case (ArrayType(e1, hasNull), e2) =>
-        TypeCoercion.findTightestCommonType(e1, e2) match {
-          case Some(dt) => Seq(ArrayType(dt, hasNull), dt)
-          case _ => Seq.empty
-        }
-      case _ => Seq.empty
-    }
-  }
+  override def prettyName: String = "array_append"
 
-  override def checkInputDataTypes(): TypeCheckResult = {
-    (left.dataType, right.dataType) match {
-      case (ArrayType(e1, _), e2) if DataTypeUtils.sameType(e1, e2) => TypeCheckResult.TypeCheckSuccess
-      case (ArrayType(e1, _), e2) => DataTypeMismatch(
-        errorSubClass = "ARRAY_FUNCTION_DIFF_TYPES",
-        messageParameters = Map(
-          "functionName" -> toSQLId(prettyName),
-          "leftType" -> toSQLType(left.dataType),
-          "rightType" -> toSQLType(right.dataType),
-          "dataType" -> toSQLType(ArrayType)
-        ))
-      case _ =>
-        DataTypeMismatch(
-          errorSubClass = "UNEXPECTED_INPUT_TYPE",
-          messageParameters = Map(
-            "paramIndex" -> "0",
-            "requiredType" -> toSQLType(ArrayType),
-            "inputSql" -> toSQLExpr(left),
-            "inputType" -> toSQLType(left.dataType)
-          )
-        )
-    }
+  override protected def withNewChildrenInternal(
+      newLeft: Expression, newRight: Expression): ArrayAppend = {
+    copy(left = newLeft, right = newRight)
   }
+}
 
-  override def eval(input: InternalRow): Any = {
-    val value1 = left.eval(input)
-    if (value1 == null) {
-      null
-    } else {
-      val value2 = right.eval(input)
-      nullSafeEval(value1, value2)
-    }
-  }
+@ExpressionDescription(
+  usage = """
+      _FUNC_(array, element) - Add the element at the beginning of the array passed as first
+      argument. Type of element should be the same as the type of the elements of the array.
+      Null element is also prepended to the array. But if the array passed is NULL
+      output is NULL
+    """,
+  examples = """
+    Examples:
+      > SELECT _FUNC_(array('b', 'd', 'c', 'a'), 'd');
+       ["d","b","d","c","a"]
+      > SELECT _FUNC_(array(1, 2, 3, null), null);
+       [null,1,2,3,null]
+      > SELECT _FUNC_(CAST(null as Array<Int>), 2);
+       NULL
+  """,
+  group = "array_funcs",
+  since = "3.5.0")
+case class ArrayPrepend(left: Expression, right: Expression) extends InsertArrayOneSide {
 
-  override protected def nullSafeEval(arr: Any, elementData: Any): Any = {
-    val arrayData = arr.asInstanceOf[ArrayData]
-    val numberOfElements = arrayData.numElements() + 1
-    if (numberOfElements > ByteArrayMethods.MAX_ROUNDED_ARRAY_LENGTH) {
-      throw QueryExecutionErrors.concatArraysWithElementsExceedLimitError(numberOfElements)
-    }
-    val finalData = new Array[Any](numberOfElements)
-    arrayData.foreach(elementType, finalData.update)
-    finalData.update(arrayData.numElements(), elementData)
-    new GenericArrayData(finalData)
-  }
+  override lazy val replacement: Expression = ArrayInsert(left, Literal(0), right)

Review Comment:
   @cloud-fan it is still 1-based, but also treat 0 as the first item. see the discussion https://github.com/apache/spark/pull/38867#discussion_r1093243515
   
   ```
   +----------------------------------+----------------------------------+----------------------------------+
   |array_insert(array(1, 2, 3), 1, 4)|array_insert(array(1, 2, 3), 0, 4)|array_insert(array(1, 2, 3), 2, 4)|
   +----------------------------------+----------------------------------+----------------------------------+
   |                      [4, 1, 2, 3]|                      [4, 1, 2, 3]|                      [1, 4, 2, 3]|
   +----------------------------------+----------------------------------+----------------------------------+
   ```



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org