MaxGekk commented on a change in pull request #28057: [SPARK-31296][SQL][TESTS] Benchmark date-time rebasing in Parquet datasource URL: https://github.com/apache/spark/pull/28057#discussion_r399931369
########## File path: sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/DateTimeRebaseBenchmark.scala ########## @@ -0,0 +1,161 @@ +/* + * 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.execution.benchmark + +import java.time.{LocalDate, LocalDateTime, LocalTime, ZoneOffset} + +import org.apache.spark.benchmark.Benchmark +import org.apache.spark.sql.DataFrame +import org.apache.spark.sql.catalyst.util.DateTimeConstants.SECONDS_PER_DAY +import org.apache.spark.sql.internal.SQLConf + +/** + * Synthetic benchmark for rebasing of date and timestamp in read/write. + * To run this benchmark: + * {{{ + * 1. without sbt: + * bin/spark-submit --class <this class> --jars <spark core test jar> <sql core test jar> + * 2. build/sbt "sql/test:runMain <this class>" + * 3. generate result: + * SPARK_GENERATE_BENCHMARK_FILES=1 build/sbt "sql/test:runMain <this class>" + * Results will be written to "benchmarks/DateTimeRebaseBenchmark-results.txt". + * }}} + */ +object DateTimeRebaseBenchmark extends SqlBasedBenchmark { + import spark.implicits._ + + private def genTs(cardinality: Int, start: LocalDateTime, end: LocalDateTime): DataFrame = { + val startSec = start.toEpochSecond(ZoneOffset.UTC) + val endSec = end.toEpochSecond(ZoneOffset.UTC) + spark.range(0, cardinality, 1, 1) + .select((($"id" % (endSec - startSec)) + startSec).as("seconds")) + .select($"seconds".cast("timestamp").as("ts")) + } + + private def genTsAfter1582(cardinality: Int): DataFrame = { + val start = LocalDateTime.of(1582, 10, 15, 0, 0, 0) + val end = LocalDateTime.of(3000, 1, 1, 0, 0, 0) + genTs(cardinality, start, end) + } + + private def genTsBefore1582(cardinality: Int): DataFrame = { + val start = LocalDateTime.of(10, 1, 1, 0, 0, 0) + val end = LocalDateTime.of(1580, 1, 1, 0, 0, 0) + genTs(cardinality, start, end) + } + + private def genDate(cardinality: Int, start: LocalDate, end: LocalDate): DataFrame = { + val startSec = LocalDateTime.of(start, LocalTime.MIDNIGHT).toEpochSecond(ZoneOffset.UTC) + val endSec = LocalDateTime.of(end, LocalTime.MIDNIGHT).toEpochSecond(ZoneOffset.UTC) + spark.range(0, cardinality * SECONDS_PER_DAY, SECONDS_PER_DAY, 1) + .select((($"id" % (endSec - startSec)) + startSec).as("seconds")) + .select($"seconds".cast("timestamp").as("ts")) + .select($"ts".cast("date").as("date")) + } + + private def genDateAfter1582(cardinality: Int): DataFrame = { + val start = LocalDate.of(1582, 10, 15) + val end = LocalDate.of(3000, 1, 1) + genDate(cardinality, start, end) + } + + private def genDateBefore1582(cardinality: Int): DataFrame = { + val start = LocalDate.of(10, 1, 1) + val end = LocalDate.of(1580, 1, 1) + genDate(cardinality, start, end) + } + + private def genDF(cardinality: Int, dateTime: String, after1582: Boolean): DataFrame = { + (dateTime, after1582) match { + case ("date", true) => genDateAfter1582(cardinality) + case ("date", false) => genDateBefore1582(cardinality) + case ("timestamp", true) => genTsAfter1582(cardinality) + case ("timestamp", false) => genTsBefore1582(cardinality) + case _ => throw new IllegalArgumentException( + s"cardinality = $cardinality dateTime = $dateTime after1582 = $after1582") + } + } + + override def runBenchmarkSuite(mainArgs: Array[String]): Unit = { + withTempPath { path => + runBenchmark("Parquet read/write") { + val rowsNum = 100000000 + Seq("date", "timestamp").foreach { dateTime => + val benchmark = new Benchmark(s"Save ${dateTime}s to parquet", rowsNum, output = output) + benchmark.addCase("after 1582, noop", 1) { _ => + genDF(rowsNum, dateTime, after1582 = true).noop() Review comment: For example: ``` after 1582, noop 9272 9272 0 10.8 92.7 1.0X ``` ``` after 1582, rebase off 21841 21841 0 4.6 218.4 0.4X ``` The `noop` benchmark shows non-avoidable overhead. If we subtract it, we get 21841 - 9272 = 12569. So, overhead of preparing input data is roughly 45%. I do believe this is important info, and we should keep in the benchmark results. ---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
