cloud-fan commented on a change in pull request #25024: [SPARK-27296][SQL] 
Allows Aggregator to be registered as a UDF
URL: https://github.com/apache/spark/pull/25024#discussion_r363161409
 
 

 ##########
 File path: 
sql/core/src/main/scala/org/apache/spark/sql/execution/aggregate/udaf.scala
 ##########
 @@ -450,3 +454,88 @@ case class ScalaUDAF(
 
   override def nodeName: String = udaf.getClass.getSimpleName
 }
+
+case class ScalaAggregator[IN, BUF, OUT](
+    children: Seq[Expression],
+    agg: Aggregator[IN, BUF, OUT],
+    inputEncoderNR: ExpressionEncoder[IN],
+    isNullable: Boolean = true,
+    isDeterministic: Boolean = true,
+    mutableAggBufferOffset: Int = 0,
+    inputAggBufferOffset: Int = 0)
+  extends TypedImperativeAggregate[BUF]
+  with NonSQLExpression
+  with UserDefinedExpression
+  with ImplicitCastInputTypes
+  with Logging {
+
+  private[this] lazy val inputEncoder = inputEncoderNR.resolveAndBind()
+  private[this] lazy val bufferEncoder = 
agg.bufferEncoder.asInstanceOf[ExpressionEncoder[BUF]]
+  private[this] lazy val outputEncoder = 
agg.outputEncoder.asInstanceOf[ExpressionEncoder[OUT]]
+
+  def dataType: DataType = outputEncoder.objSerializer.dataType
+
+  def inputTypes: Seq[DataType] = inputEncoder.schema.map(_.dataType)
+
+  def nullable: Boolean = isNullable
+
+  override lazy val deterministic: Boolean = isDeterministic
+
+  def withNewMutableAggBufferOffset(newMutableAggBufferOffset: Int): 
ScalaAggregator[IN, BUF, OUT] =
+    copy(mutableAggBufferOffset = newMutableAggBufferOffset)
+
+  def withNewInputAggBufferOffset(newInputAggBufferOffset: Int): 
ScalaAggregator[IN, BUF, OUT] =
+    copy(inputAggBufferOffset = newInputAggBufferOffset)
+
+  private[this] lazy val childrenSchema: StructType = {
+    val inputFields = children.zipWithIndex.map {
+      case (child, index) =>
+        StructField(s"input$index", child.dataType, child.nullable, 
Metadata.empty)
+    }
+    StructType(inputFields)
+  }
+
+  private[this] lazy val inputProjection = {
+    val inputAttributes = childrenSchema.toAttributes
+    log.debug(
+      s"Creating MutableProj: $children, inputSchema: $inputAttributes.")
+    UnsafeProjection.create(children, inputAttributes)
+  }
+
+  def createAggregationBuffer(): BUF = agg.zero
+
+  def update(buffer: BUF, input: InternalRow): BUF = {
+    val proj = inputProjection(input)
+    val a = inputEncoder.fromRow(proj)
+    agg.reduce(buffer, a)
 
 Review comment:
   nit: we can put them in one line
   `agg.reduce(buffer, inputEncoder.fromRow(inputProjection(input)))`
   
   to avoid unclear naming like `proj`, `a`.

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


With regards,
Apache Git Services

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

Reply via email to