[GitHub] [spark] maropu commented on a change in pull request #32243: [WIP][SPARK-35192][SQL][TESTS] Port minimal TPC-DS datagen code from databricks/spark-sql-perf

[GitBox <gi...@apache.org>]

maropu commented on a change in pull request #32243:
URL: https://github.com/apache/spark/pull/32243#discussion_r623872242



##########
File path: sql/core/src/test/scala/org/apache/spark/sql/tpcdsDagagen.scala
##########
@@ -0,0 +1,443 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements.  See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License.  You may obtain a copy of the License at
+ *
+ *    http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.sql
+
+import java.util.concurrent.LinkedBlockingQueue
+
+import scala.collection.immutable.Stream
+import scala.sys.process._
+import scala.util.Try
+
+import org.apache.spark.SparkContext
+import org.apache.spark.internal.Logging
+import org.apache.spark.rdd.RDD
+import org.apache.spark.sql.functions.{col, rpad}
+import org.apache.spark.sql.types.{CharType, StringType, StructField, StructType, VarcharType}
+
+// The classes in this file are basically moved from https://github.com/databricks/spark-sql-perf
+
+/**
+ * Using ProcessBuilder.lineStream produces a stream, that uses
+ * a LinkedBlockingQueue with a default capacity of Integer.MAX_VALUE.
+ *
+ * This causes OOM if the consumer cannot keep up with the producer.
+ *
+ * See scala.sys.process.ProcessBuilderImpl.lineStream
+ */
+object BlockingLineStream {
+
+  // See scala.sys.process.Streamed
+  private final class BlockingStreamed[T](
+    val process: T => Unit,
+    val done: Int => Unit,
+    val stream: () => Stream[T])
+
+  // See scala.sys.process.Streamed
+  private object BlockingStreamed {
+    // scala.process.sys.Streamed uses default of Integer.MAX_VALUE,
+    // which causes OOMs if the consumer cannot keep up with producer.
+    val maxQueueSize = 65536
+
+    def apply[T](nonzeroException: Boolean): BlockingStreamed[T] = {
+      val q = new LinkedBlockingQueue[Either[Int, T]](maxQueueSize)
+
+      def next(): Stream[T] = q.take match {
+        case Left(0) => Stream.empty
+        case Left(code) =>
+          if (nonzeroException) scala.sys.error("Nonzero exit code: " + code) else Stream.empty
+        case Right(s) => Stream.cons(s, next())
+      }
+
+      new BlockingStreamed((s: T) => q put Right(s), code => q put Left(code), () => next())
+    }
+  }
+
+  // See scala.sys.process.ProcessImpl.Spawn
+  private object Spawn {
+    def apply(f: => Unit): Thread = apply(f, daemon = false)
+    def apply(f: => Unit, daemon: Boolean): Thread = {
+      val thread = new Thread() { override def run() = { f } }
+      thread.setDaemon(daemon)
+      thread.start()
+      thread
+    }
+  }
+
+  def apply(command: Seq[String]): Stream[String] = {
+    val streamed = BlockingStreamed[String](true)
+    val process = command.run(BasicIO(false, streamed.process, None))
+    Spawn(streamed.done(process.exitValue()))
+    streamed.stream()
+  }
+}
+
+class Dsdgen(dsdgenDir: String) extends Serializable {
+  private val dsdgen = s"$dsdgenDir/dsdgen"
+
+  def generate(
+      sparkContext: SparkContext,
+      tableName: String,
+      partitions: Int,
+      scaleFactor: Int): RDD[String] = {
+    val generatedData = {
+      sparkContext.parallelize(1 to partitions, partitions).flatMap { i =>
+        val localToolsDir = if (new java.io.File(dsdgen).exists) {
+          dsdgenDir
+        } else if (new java.io.File(s"/$dsdgen").exists) {
+          s"/$dsdgenDir"
+        } else {
+          throw new IllegalStateException(
+            s"Could not find dsdgen at $dsdgen or /$dsdgen. Run install")
+        }
+
+        // NOTE: RNGSEED is the RNG seed used by the data generator. Right now, it is fixed to
+        // 19620718 that is used to generate `https://github.com/maropu/spark-tpcds-sf-1`.
+        val parallel = if (partitions > 1) s"-parallel $partitions -child $i" else ""
+        val commands = Seq(
+          "bash", "-c",
+          s"cd $localToolsDir && ./dsdgen -table $tableName -filter Y -scale $scaleFactor " +
+          s"-RNGSEED 19620718 $parallel")

Review comment:
       NOTE: In a follow-up PR, I'll revert back this value to 100 to use `https://github.com/databricks/tpcds-kit` instead. See: https://github.com/apache/spark/pull/32243#issuecomment-824827326




-- 
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:
users@infra.apache.org



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