cloud-fan

object NewUDFBenchmark extends SqlBasedBenchmark {
  import spark.implicits._

  private def nativeAdd(card: Long): Unit = {
    spark.range(card).select($"id" + $"id").write.format("noop").mode("append").save()
  }

  private def udfAdd(card: Long): Unit = {
    val my_udf = udf { (input1: Long, input2: Long) => input1 + input2 }
    spark.range(card).select(my_udf($"id", $"id")).write.format("noop").mode("append").save()

cloud-fan

normal table: src(a int, b int)

the result is produced by select a, b from src

SQL:

INSERT INTO TABLE src SELECT 1,2;

1.6:

• hive table: success, the result is 1, 2

cloud-fan

object ArrayHashBenchmark {

  def main (args: Array[String]) {
    val iters = 1024 * 40
    val numArrays = 1024

    val schema = ArrayType(IntegerType)
    val generator = RandomDataGenerator.forType(schema, nullable = false).get
    val encoder = ExpressionEncoder[Seq[Int]]

cloud-fan

import scala.collection.mutable

class CostFlow(src: String, target: String) {
  outer =>

  class Edge(
      val to: String,
      val cost: Int,
      var capacity: Int,
      val next: Int,

cloud-fan

import scala.collection.mutable
import scala.collection.immutable.HashMap

class Graph(data: Seq[(String, String, Double)]) {

  case class Vertex(label: String, connection: Seq[Edge])

  class Edge(targetLabel: String, val weight: Double) {
    lazy val target = adjacencyList(targetLabel)
  }