Github user rxin commented on a diff in the pull request:
https://github.com/apache/spark/pull/12087#discussion_r58646052
--- Diff:
sql/core/src/test/scala/org/apache/spark/sql/DatasetBenchmark.scala ---
@@ -0,0 +1,85 @@
+/*
+ * 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 org.apache.spark.SparkContext
+import org.apache.spark.sql.types.StringType
+import org.apache.spark.util.Benchmark
+
+/**
+ * Benchmark for Dataset typed operations comparing with DataFrame and RDD
versions.
+ */
+object DatasetBenchmark {
+
+ case class Data(l: Long, s: String)
+
+ def main(args: Array[String]): Unit = {
+ val sparkContext = new SparkContext("local[*]", "Dataset benchmark")
+ val sqlContext = new SQLContext(sparkContext)
+
+ import sqlContext.implicits._
+
+ val numRows = 10000000
+ val df = sqlContext.range(1, numRows).select($"id".as("l"),
$"id".cast(StringType).as("s"))
+
+ val benchmark = new Benchmark("back-to-back map", numRows)
+
+ val scalaFunc = (d: Data) => Data(d.l + 1, d.s)
+ benchmark.addCase("Dataset") { iter =>
+ var res = df.as[Data]
+ var i = 0
+ while (i < 10) {
+ res = res.map(scalaFunc)
+ i += 1
+ }
+ res.queryExecution.toRdd.foreach(_ => Unit)
+ }
+
+ benchmark.addCase("DataFrame") { iter =>
+ var res = df
+ var i = 0
+ while (i < 10) {
+ res = res.select($"l" + 1 as "l")
+ i += 1
+ }
+ res.queryExecution.toRdd.foreach(_ => Unit)
+ }
+
+ val rdd = sparkContext.range(1, numRows).map(l => Data(l, l.toString))
+ benchmark.addCase("RDD") { iter =>
+ var res = rdd
+ var i = 0
+ while (i < 10) {
+ res = rdd.map(scalaFunc)
+ i += 1
+ }
+ res.foreach(_ => Unit)
+ }
+
+ /*
+ Java HotSpot(TM) 64-Bit Server VM 1.8.0_60-b27 on Mac OS X 10.11.4
+ Intel(R) Core(TM) i7-4960HQ CPU @ 2.60GHz
+ back-to-back map: Best/Avg Time(ms) Rate(M/s)
Per Row(ns) Relative
+
-------------------------------------------------------------------------------------------
+ Dataset 902 / 995 11.1
90.2 1.0X
+ DataFrame 132 / 167 75.5
13.2 6.8X
+ RDD 216 / 237 46.3
21.6 4.2X
--- End diff --
"And the reason that you would make a case class mutable other than to
allow reuse is"
because you might want to mutate state? this shouldn't work only for case
classes -- there are other things we should reuse, e.g. tuples, java beans. I
wouldn't want to overload an existing language feature that is not about object
reuse.
I do agree it'd be better to reuse by default for performance improvements,
and users can explicitly turn it off through a flag.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
