beliefer commented on code in PR #41856:
URL: https://github.com/apache/spark/pull/41856#discussion_r1257098969
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCQueryBenchmarkArguments.scala:
##########
@@ -63,15 +63,15 @@ class TPCDSQueryBenchmarkArguments(val args: Array[String])
{
System.err.println("""
|Usage: spark-submit --class <this class> <spark sql test jar> [Options]
|Options:
- | --data-location Path to TPCDS data
+ | --data-location Path to TPCDS/H data
| --query-filter Queries to filter, e.g., q3,q5,q13
| --cbo Whether to enable cost-based optimization
|
|------------------------------------------------------------------------------------------------------------------
|In order to run this benchmark, please follow the instructions at
|https://github.com/databricks/spark-sql-perf/blob/master/README.md
- |to generate the TPCDS data locally (preferably with a scale factor of 5
for benchmarking).
- |Thereafter, the value of <TPCDS data location> needs to be set to the
location where the generated data is stored.
+ |to generate the TPCDS/H data locally (preferably with a scale factor of
5 for benchmarking).
Review Comment:
It seems a little odd. May it be `TPC-DS/TPC-H`
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCQueryBenchmarkArguments.scala:
##########
@@ -63,15 +63,15 @@ class TPCDSQueryBenchmarkArguments(val args: Array[String])
{
System.err.println("""
|Usage: spark-submit --class <this class> <spark sql test jar> [Options]
|Options:
- | --data-location Path to TPCDS data
+ | --data-location Path to TPCDS/H data
| --query-filter Queries to filter, e.g., q3,q5,q13
| --cbo Whether to enable cost-based optimization
|
|------------------------------------------------------------------------------------------------------------------
|In order to run this benchmark, please follow the instructions at
|https://github.com/databricks/spark-sql-perf/blob/master/README.md
- |to generate the TPCDS data locally (preferably with a scale factor of 5
for benchmarking).
- |Thereafter, the value of <TPCDS data location> needs to be set to the
location where the generated data is stored.
+ |to generate the TPCDS/H data locally (preferably with a scale factor of
5 for benchmarking).
+ |Thereafter, the value of <TPCDS/H data location> needs to be set to the
location where the generated data is stored.
Review Comment:
ditto.
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCBasedBenchmark.scala:
##########
@@ -0,0 +1,134 @@
+/*
+ * 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 scala.util.Try
+
+import org.apache.spark.SparkConf
+import org.apache.spark.benchmark.Benchmark
+import org.apache.spark.internal.Logging
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.catalog.HiveTableRelation
+import org.apache.spark.sql.catalyst.plans.logical.SubqueryAlias
+import org.apache.spark.sql.catalyst.util.DateTimeConstants.NANOS_PER_SECOND
+import org.apache.spark.sql.catalyst.util.resourceToString
+import org.apache.spark.sql.execution.datasources.LogicalRelation
+import org.apache.spark.sql.internal.SQLConf
+import org.apache.spark.sql.types.StructType
+
+
+trait TPCBasedBenchmark extends SqlBasedBenchmark with Logging {
+
+ val tables: Seq[String]
+ val queryType: String
+
+ override def getSparkSession: SparkSession = {
+ val conf = new SparkConf()
+ .setMaster(System.getProperty("spark.sql.test.master", "local[1]"))
+ .setAppName("test-sql-context")
+ .set("spark.sql.parquet.compression.codec", "snappy")
+ .set("spark.sql.shuffle.partitions",
System.getProperty("spark.sql.shuffle.partitions", "4"))
+ .set("spark.driver.memory", "3g")
+ .set("spark.executor.memory", "3g")
+ .set("spark.sql.autoBroadcastJoinThreshold", (20 * 1024 * 1024).toString)
+ .set("spark.sql.crossJoin.enabled", "true")
+ .set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
+ .set("spark.kryo.registrationRequired", "true")
+
+ SparkSession.builder.config(conf).getOrCreate()
+ }
+
+ def setupTables(dataLocation: String, tableColumns: Map[String,
StructType]): Map[String, Long] =
+ tables.map { tableName =>
+ spark.sql(s"DROP TABLE IF EXISTS $tableName")
+ val options = Map("path" -> s"$dataLocation/$tableName")
+ spark.catalog.createTable(tableName, "parquet", tableColumns(tableName),
options)
+ // Recover partitions but don't fail if a table is not partitioned.
+ Try {
+ spark.sql(s"ALTER TABLE $tableName RECOVER PARTITIONS")
+ }.getOrElse {
+ logInfo(s"Recovering partitions of table $tableName failed")
+ }
+ tableName -> spark.table(tableName).count()
+ }.toMap
+
+ def runTpcQueries(
+ queryLocation: String,
+ queries: Seq[String],
+ tableSizes: Map[String, Long],
+ nameSuffix: String = ""): Unit = {
+ queries.foreach { name =>
+ val queryString = resourceToString(s"$queryLocation/$name.sql",
+ classLoader = Thread.currentThread().getContextClassLoader)
+
+ // This is an indirect hack to estimate the size of each query's input
by traversing the
+ // logical plan and adding up the sizes of all tables that appear in the
plan.
+ val queryRelations = scala.collection.mutable.HashSet[String]()
+ spark.sparkContext.setJobGroup(name, s"$name:\n$queryString", true)
+ spark.sql(queryString).queryExecution.analyzed.foreach {
+ case SubqueryAlias(alias, _: LogicalRelation) =>
+ queryRelations.add(alias.name)
+ case LogicalRelation(_, _, Some(catalogTable), _) =>
+ queryRelations.add(catalogTable.identifier.table)
+ case HiveTableRelation(tableMeta, _, _, _, _) =>
+ queryRelations.add(tableMeta.identifier.table)
+ case _ =>
+ }
+ val numRows = queryRelations.map(tableSizes.getOrElse(_, 0L)).sum
+ val benchmark = new Benchmark(s"$queryType Snappy", numRows, 2, output =
output)
+ benchmark.addCase(s"$name$nameSuffix") { _ =>
+ spark.sql(queryString).noop()
+ }
+ benchmark.run()
+ }
+ }
+
+ def filterQueries(
Review Comment:
ditto
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCBasedBenchmark.scala:
##########
@@ -0,0 +1,134 @@
+/*
+ * 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 scala.util.Try
+
+import org.apache.spark.SparkConf
+import org.apache.spark.benchmark.Benchmark
+import org.apache.spark.internal.Logging
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.catalog.HiveTableRelation
+import org.apache.spark.sql.catalyst.plans.logical.SubqueryAlias
+import org.apache.spark.sql.catalyst.util.DateTimeConstants.NANOS_PER_SECOND
+import org.apache.spark.sql.catalyst.util.resourceToString
+import org.apache.spark.sql.execution.datasources.LogicalRelation
+import org.apache.spark.sql.internal.SQLConf
+import org.apache.spark.sql.types.StructType
+
+
+trait TPCBasedBenchmark extends SqlBasedBenchmark with Logging {
+
+ val tables: Seq[String]
+ val queryType: String
+
+ override def getSparkSession: SparkSession = {
+ val conf = new SparkConf()
+ .setMaster(System.getProperty("spark.sql.test.master", "local[1]"))
+ .setAppName("test-sql-context")
+ .set("spark.sql.parquet.compression.codec", "snappy")
+ .set("spark.sql.shuffle.partitions",
System.getProperty("spark.sql.shuffle.partitions", "4"))
+ .set("spark.driver.memory", "3g")
+ .set("spark.executor.memory", "3g")
+ .set("spark.sql.autoBroadcastJoinThreshold", (20 * 1024 * 1024).toString)
+ .set("spark.sql.crossJoin.enabled", "true")
+ .set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
+ .set("spark.kryo.registrationRequired", "true")
+
+ SparkSession.builder.config(conf).getOrCreate()
+ }
+
+ def setupTables(dataLocation: String, tableColumns: Map[String,
StructType]): Map[String, Long] =
+ tables.map { tableName =>
+ spark.sql(s"DROP TABLE IF EXISTS $tableName")
+ val options = Map("path" -> s"$dataLocation/$tableName")
+ spark.catalog.createTable(tableName, "parquet", tableColumns(tableName),
options)
+ // Recover partitions but don't fail if a table is not partitioned.
+ Try {
+ spark.sql(s"ALTER TABLE $tableName RECOVER PARTITIONS")
+ }.getOrElse {
+ logInfo(s"Recovering partitions of table $tableName failed")
+ }
+ tableName -> spark.table(tableName).count()
+ }.toMap
+
+ def runTpcQueries(
Review Comment:
ditto.
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCQueryBenchmarkArguments.scala:
##########
@@ -63,15 +63,15 @@ class TPCDSQueryBenchmarkArguments(val args: Array[String])
{
System.err.println("""
|Usage: spark-submit --class <this class> <spark sql test jar> [Options]
|Options:
- | --data-location Path to TPCDS data
+ | --data-location Path to TPCDS/H data
Review Comment:
ditto
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCBasedBenchmark.scala:
##########
@@ -0,0 +1,134 @@
+/*
+ * 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 scala.util.Try
+
+import org.apache.spark.SparkConf
+import org.apache.spark.benchmark.Benchmark
+import org.apache.spark.internal.Logging
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.catalog.HiveTableRelation
+import org.apache.spark.sql.catalyst.plans.logical.SubqueryAlias
+import org.apache.spark.sql.catalyst.util.DateTimeConstants.NANOS_PER_SECOND
+import org.apache.spark.sql.catalyst.util.resourceToString
+import org.apache.spark.sql.execution.datasources.LogicalRelation
+import org.apache.spark.sql.internal.SQLConf
+import org.apache.spark.sql.types.StructType
+
+
+trait TPCBasedBenchmark extends SqlBasedBenchmark with Logging {
+
+ val tables: Seq[String]
+ val queryType: String
+
+ override def getSparkSession: SparkSession = {
+ val conf = new SparkConf()
+ .setMaster(System.getProperty("spark.sql.test.master", "local[1]"))
+ .setAppName("test-sql-context")
+ .set("spark.sql.parquet.compression.codec", "snappy")
+ .set("spark.sql.shuffle.partitions",
System.getProperty("spark.sql.shuffle.partitions", "4"))
+ .set("spark.driver.memory", "3g")
+ .set("spark.executor.memory", "3g")
+ .set("spark.sql.autoBroadcastJoinThreshold", (20 * 1024 * 1024).toString)
+ .set("spark.sql.crossJoin.enabled", "true")
+ .set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
+ .set("spark.kryo.registrationRequired", "true")
+
+ SparkSession.builder.config(conf).getOrCreate()
+ }
+
+ def setupTables(dataLocation: String, tableColumns: Map[String,
StructType]): Map[String, Long] =
Review Comment:
```suggestion
protected def setupTables(dataLocation: String, tableColumns: Map[String,
StructType]): Map[String, Long] =
```
##########
sql/core/src/test/scala/org/apache/spark/sql/execution/benchmark/TPCBasedBenchmark.scala:
##########
@@ -0,0 +1,134 @@
+/*
+ * 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 scala.util.Try
+
+import org.apache.spark.SparkConf
+import org.apache.spark.benchmark.Benchmark
+import org.apache.spark.internal.Logging
+import org.apache.spark.sql.SparkSession
+import org.apache.spark.sql.catalyst.catalog.HiveTableRelation
+import org.apache.spark.sql.catalyst.plans.logical.SubqueryAlias
+import org.apache.spark.sql.catalyst.util.DateTimeConstants.NANOS_PER_SECOND
+import org.apache.spark.sql.catalyst.util.resourceToString
+import org.apache.spark.sql.execution.datasources.LogicalRelation
+import org.apache.spark.sql.internal.SQLConf
+import org.apache.spark.sql.types.StructType
+
+
+trait TPCBasedBenchmark extends SqlBasedBenchmark with Logging {
+
+ val tables: Seq[String]
+ val queryType: String
+
+ override def getSparkSession: SparkSession = {
+ val conf = new SparkConf()
+ .setMaster(System.getProperty("spark.sql.test.master", "local[1]"))
+ .setAppName("test-sql-context")
+ .set("spark.sql.parquet.compression.codec", "snappy")
+ .set("spark.sql.shuffle.partitions",
System.getProperty("spark.sql.shuffle.partitions", "4"))
+ .set("spark.driver.memory", "3g")
+ .set("spark.executor.memory", "3g")
+ .set("spark.sql.autoBroadcastJoinThreshold", (20 * 1024 * 1024).toString)
+ .set("spark.sql.crossJoin.enabled", "true")
+ .set("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
+ .set("spark.kryo.registrationRequired", "true")
+
+ SparkSession.builder.config(conf).getOrCreate()
+ }
+
+ def setupTables(dataLocation: String, tableColumns: Map[String,
StructType]): Map[String, Long] =
+ tables.map { tableName =>
+ spark.sql(s"DROP TABLE IF EXISTS $tableName")
+ val options = Map("path" -> s"$dataLocation/$tableName")
+ spark.catalog.createTable(tableName, "parquet", tableColumns(tableName),
options)
+ // Recover partitions but don't fail if a table is not partitioned.
+ Try {
+ spark.sql(s"ALTER TABLE $tableName RECOVER PARTITIONS")
+ }.getOrElse {
+ logInfo(s"Recovering partitions of table $tableName failed")
+ }
+ tableName -> spark.table(tableName).count()
+ }.toMap
+
+ def runTpcQueries(
+ queryLocation: String,
+ queries: Seq[String],
+ tableSizes: Map[String, Long],
+ nameSuffix: String = ""): Unit = {
+ queries.foreach { name =>
+ val queryString = resourceToString(s"$queryLocation/$name.sql",
+ classLoader = Thread.currentThread().getContextClassLoader)
+
+ // This is an indirect hack to estimate the size of each query's input
by traversing the
+ // logical plan and adding up the sizes of all tables that appear in the
plan.
+ val queryRelations = scala.collection.mutable.HashSet[String]()
+ spark.sparkContext.setJobGroup(name, s"$name:\n$queryString", true)
+ spark.sql(queryString).queryExecution.analyzed.foreach {
+ case SubqueryAlias(alias, _: LogicalRelation) =>
+ queryRelations.add(alias.name)
+ case LogicalRelation(_, _, Some(catalogTable), _) =>
+ queryRelations.add(catalogTable.identifier.table)
+ case HiveTableRelation(tableMeta, _, _, _, _) =>
+ queryRelations.add(tableMeta.identifier.table)
+ case _ =>
+ }
+ val numRows = queryRelations.map(tableSizes.getOrElse(_, 0L)).sum
+ val benchmark = new Benchmark(s"$queryType Snappy", numRows, 2, output =
output)
+ benchmark.addCase(s"$name$nameSuffix") { _ =>
+ spark.sql(queryString).noop()
+ }
+ benchmark.run()
+ }
+ }
+
+ def filterQueries(
+ origQueries: Seq[String],
+ queryFilter: Set[String],
+ nameSuffix: String = ""): Seq[String] = {
+ if (queryFilter.nonEmpty) {
+ if (nameSuffix.nonEmpty) {
+ origQueries.filter { name => queryFilter.contains(s"$name$nameSuffix")
}
+ } else {
+ origQueries.filter(queryFilter.contains)
+ }
+ } else {
+ origQueries
+ }
+ }
+
+ def configureCbo(benchmarkArgs: TPCQueryBenchmarkArguments): Unit = {
Review Comment:
ditto
--
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.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
