apache/kyuubi
1
0
Fork 0
Code Issues Pull Requests Actions 5 Packages Projects Releases Wiki Activity

[KYUUBI #6857] Spark 3.4: MaxScanStrategy supports DSv2

Browse Source 
### Why are the changes needed?

Backport https://github.com/apache/kyuubi/pull/5852 to Spark 3.4, to enhance MaxScanStrategy to include support for the datasourcev2 in Spark 3.4

### How was this patch tested?

Add some UTs

### Was this patch authored or co-authored using generative AI tooling?

No

Closes #6857 from zhaohehuhu/dev-1224.

Closes #6857

c72c62984 [zhaohehuhu] remove the import
dfbf2bc2d [zhaohehuhu] MaxScanStrategy supports DSv2 in Spark 3.4

Authored-by: zhaohehuhu <luoyedeyi459@163.com>
Signed-off-by: Cheng Pan <chengpan@apache.org>
This commit is contained in:
zhaohehuhu 2024-12-24 20:24:03 +08:00 committed by Cheng Pan
parent 7e8275b7b4
commit 3c72fef476
No known key found for this signature in database
GPG Key ID: 8001952629BCC75D
4 changed files with 182 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
extensions/spark/kyuubi-extension-spark-3-4/src/main/scala/org/apache/kyuubi/sql/watchdog/MaxScanStrategy.scala
View File

@ -25,6 +25,7 @@ import org.apache.spark.sql.catalyst.planning.ScanOperation
import org.apache.spark.sql.catalyst.plans.logical.LogicalPlan
import org.apache.spark.sql.execution.SparkPlan
import org.apache.spark.sql.execution.datasources.{CatalogFileIndex, HadoopFsRelation, InMemoryFileIndex, LogicalRelation}
import org.apache.spark.sql.execution.datasources.v2.DataSourceV2ScanRelation
import org.apache.spark.sql.types.StructType
import org.apache.kyuubi.sql.KyuubiSQLConf
@ -232,6 +233,40 @@ case class MaxScanStrategy(session: SparkSession)
logicalRelation.catalogTable)
}
}
case ScanOperation(
_,
_,
_,
relation @ DataSourceV2ScanRelation(_, _, _, _, _)) =>
val table = relation.relation.table
if (table.partitioning().nonEmpty) {
val partitionColumnNames = table.partitioning().map(_.describe())
val stats = relation.computeStats()
lazy val scanFileSize = stats.sizeInBytes
if (maxFileSizeOpt.exists(_ < scanFileSize)) {
throw new MaxFileSizeExceedException(
s"""
|SQL job scan file size in bytes: $scanFileSize
|exceed restrict of table scan maxFileSize ${maxFileSizeOpt.get}
|You should optimize your SQL logical according partition structure
|or shorten query scope such as p_date, detail as below:
|Table: ${table.name()}
|Partition Structure: ${partitionColumnNames.mkString(",")}
|""".stripMargin)
}
} else {
val stats = relation.computeStats()
lazy val scanFileSize = stats.sizeInBytes
if (maxFileSizeOpt.exists(_ < scanFileSize)) {
throw new MaxFileSizeExceedException(
s"""
|SQL job scan file size in bytes: $scanFileSize
|exceed restrict of table scan maxFileSize ${maxFileSizeOpt.get}
|detail as below:
|Table: ${table.name()}
|""".stripMargin)
}
}
case _ =>
}
}

64
extensions/spark/kyuubi-extension-spark-3-4/src/test/scala/org/apache/spark/sql/ReportStatisticsAndPartitionAwareDataSource.scala Normal file
View File

@ -0,0 +1,64 @@
/*
* 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.OptionalLong
import org.apache.spark.sql.connector.{RangeInputPartition, SimpleBatchTable, SimpleScanBuilder, SimpleWritableDataSource}
import org.apache.spark.sql.connector.catalog.Table
import org.apache.spark.sql.connector.expressions.{Expressions, FieldReference, Transform}
import org.apache.spark.sql.connector.read._
import org.apache.spark.sql.connector.read.partitioning.{KeyGroupedPartitioning, Partitioning}
import org.apache.spark.sql.util.CaseInsensitiveStringMap
class ReportStatisticsAndPartitionAwareDataSource extends SimpleWritableDataSource {
class MyScanBuilder(
val partitionKeys: Seq[String]) extends SimpleScanBuilder
with SupportsReportStatistics with SupportsReportPartitioning {
override def estimateStatistics(): Statistics = {
new Statistics {
override def sizeInBytes(): OptionalLong = OptionalLong.of(80)
override def numRows(): OptionalLong = OptionalLong.of(10)
}
}
override def planInputPartitions(): Array[InputPartition] = {
Array(RangeInputPartition(0, 5), RangeInputPartition(5, 10))
}
override def outputPartitioning(): Partitioning = {
new KeyGroupedPartitioning(partitionKeys.map(FieldReference(_)).toArray, 10)
}
}
override def getTable(options: CaseInsensitiveStringMap): Table = {
new SimpleBatchTable {
override def newScanBuilder(options: CaseInsensitiveStringMap): ScanBuilder = {
new MyScanBuilder(Seq("i"))
}
override def partitioning(): Array[Transform] = {
Array(Expressions.identity("i"))
}
}
}
}

53
extensions/spark/kyuubi-extension-spark-3-4/src/test/scala/org/apache/spark/sql/ReportStatisticsDataSource.scala Normal file
View File

@ -0,0 +1,53 @@
/*
* 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.OptionalLong
import org.apache.spark.sql.connector._
import org.apache.spark.sql.connector.catalog.Table
import org.apache.spark.sql.connector.read._
import org.apache.spark.sql.util.CaseInsensitiveStringMap
class ReportStatisticsDataSource extends SimpleWritableDataSource {
class MyScanBuilder extends SimpleScanBuilder
with SupportsReportStatistics {
override def estimateStatistics(): Statistics = {
new Statistics {
override def sizeInBytes(): OptionalLong = OptionalLong.of(80)
override def numRows(): OptionalLong = OptionalLong.of(10)
}
}
override def planInputPartitions(): Array[InputPartition] = {
Array(RangeInputPartition(0, 5), RangeInputPartition(5, 10))
}
}
override def getTable(options: CaseInsensitiveStringMap): Table = {
new SimpleBatchTable {
override def newScanBuilder(options: CaseInsensitiveStringMap): ScanBuilder = {
new MyScanBuilder
}
}
}
}

30
extensions/spark/kyuubi-extension-spark-3-4/src/test/scala/org/apache/spark/sql/WatchDogSuiteBase.scala
View File

@ -23,6 +23,7 @@ import scala.collection.JavaConverters._
import org.apache.commons.io.FileUtils
import org.apache.spark.sql.catalyst.plans.logical.{GlobalLimit, LogicalPlan}
import org.apache.spark.sql.execution.datasources.v2.DataSourceV2ScanRelation
import org.apache.kyuubi.sql.{KyuubiSQLConf, KyuubiSQLExtensionException}
import org.apache.kyuubi.sql.watchdog.{MaxFileSizeExceedException, MaxPartitionExceedException}
@ -607,4 +608,33 @@ trait WatchDogSuiteBase extends KyuubiSparkSQLExtensionTest {
assert(e.getMessage == "Script transformation is not allowed")
}
}
test("watchdog with scan maxFileSize -- data source v2") {
val df = spark.read.format(classOf[ReportStatisticsAndPartitionAwareDataSource].getName).load()
df.createOrReplaceTempView("test")
val logical = df.queryExecution.optimizedPlan.collect {
case d: DataSourceV2ScanRelation => d
}.head
val tableSize = logical.computeStats().sizeInBytes.toLong
withSQLConf(KyuubiSQLConf.WATCHDOG_MAX_FILE_SIZE.key -> tableSize.toString) {
sql("SELECT * FROM test").queryExecution.sparkPlan
}
withSQLConf(KyuubiSQLConf.WATCHDOG_MAX_FILE_SIZE.key -> (tableSize / 2).toString) {
intercept[MaxFileSizeExceedException](
sql("SELECT * FROM test").queryExecution.sparkPlan)
}
val nonPartDf = spark.read.format(classOf[ReportStatisticsDataSource].getName).load()
nonPartDf.createOrReplaceTempView("test_non_part")
val nonPartLogical = nonPartDf.queryExecution.optimizedPlan.collect {
case d: DataSourceV2ScanRelation => d
}.head
val nonPartTableSize = nonPartLogical.computeStats().sizeInBytes.toLong
withSQLConf(KyuubiSQLConf.WATCHDOG_MAX_FILE_SIZE.key -> nonPartTableSize.toString) {
sql("SELECT * FROM test_non_part").queryExecution.sparkPlan
}
withSQLConf(KyuubiSQLConf.WATCHDOG_MAX_FILE_SIZE.key -> (nonPartTableSize / 2).toString) {
intercept[MaxFileSizeExceedException](
sql("SELECT * FROM test_non_part").queryExecution.sparkPlan)
}
}
}