[KYUUBI #6943][1/2]HiveScan support dpp #7436
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,91 @@ | ||
| /* | ||
| * 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.kyuubi.spark.connector.hive.read | ||
|
|
||
| import java.util.Locale | ||
|
|
||
| import org.apache.spark.internal.Logging | ||
| import org.apache.spark.sql.catalyst.expressions.{AttributeReference, Expression, In, Literal} | ||
| import org.apache.spark.sql.connector.expressions.{Expressions, NamedReference} | ||
| import org.apache.spark.sql.hive.kyuubi.connector.HiveBridgeHelper.StructTypeHelper | ||
| import org.apache.spark.sql.sources.{Filter, In => FilterIn} | ||
| import org.apache.spark.sql.types.StructType | ||
|
|
||
| /** | ||
| * Helpers for a Hive-backed V2 [[org.apache.spark.sql.connector.read.Scan]] to | ||
| * implement [[org.apache.spark.sql.connector.read.SupportsRuntimeFiltering]] | ||
| * for Dynamic Partition Pruning (DPP). | ||
| * | ||
| * Spark's `DataSourceV2Strategy` currently only emits the `IN` form as a DPP | ||
| * runtime filter, so translation here handles `In` only. Any filter whose | ||
| * attribute does not match a known partition column is dropped; drops are | ||
| * logged at DEBUG. | ||
| * | ||
| * We deliberately use the V1 `SupportsRuntimeFiltering` instead of the newer | ||
| * `SupportsRuntimeV2Filtering` to keep this connector compilable against | ||
| * Spark 3.3, where `SupportsRuntimeV2Filtering` was introduced in Spark 3.4. | ||
| */ | ||
| object HiveRuntimeFilterSupport extends Logging { | ||
|
|
||
| /** | ||
| * Build the runtime-filterable attribute array. Only partition columns are exposed | ||
| * because DPP is only beneficial at the partition directory granularity. | ||
| */ | ||
| def filterAttributes(partitionColumnNames: Seq[String]): Array[NamedReference] = { | ||
| partitionColumnNames.map(Expressions.column).toArray | ||
| } | ||
|
|
||
| /** | ||
| * Translate Spark's runtime V1 `In` filters into catalyst [[In]] expressions | ||
| * bound to the given partition attributes. | ||
| * | ||
| * A filter is accepted only when it is a [[FilterIn]] whose attribute resolves | ||
| * to a known partition column. | ||
| */ | ||
| def toCatalystPartitionFilters( | ||
|
Member
There was a problem hiding this comment. This makes the code compatibility very fragile, catalyst code is treated as internal implementation details and does not provide any compatibility guarantee. Given the situation, why choose to implement
Contributor
Author
There was a problem hiding this comment. Good point. I’ll use
Contributor
Author
There was a problem hiding this comment. I've replaced |
||
| filters: Array[Filter], | ||
| partitionSchema: StructType, | ||
| isCaseSensitive: Boolean): Seq[Expression] = { | ||
| val attrByName: Map[String, AttributeReference] = | ||
| partitionSchema.toAttributes | ||
| .map(a => normalize(a.name, isCaseSensitive) -> a).toMap | ||
|
|
||
| val accepted = filters.toSeq.flatMap { | ||
| case FilterIn(attribute, values) => | ||
| attrByName.get(normalize(attribute, isCaseSensitive)).map { attr => | ||
| In(attr, values.toSeq.map(v => Literal.create(v, attr.dataType))) | ||
| } | ||
| case _ => None | ||
| } | ||
|
|
||
| if (accepted.length < filters.length) { | ||
| logDebug( | ||
| s"Dropped ${filters.length - accepted.length} of ${filters.length} runtime " + | ||
| s"filter(s) not applicable to partition columns " + | ||
| s"[${partitionSchema.fieldNames.mkString(", ")}]") | ||
| } | ||
| accepted | ||
| } | ||
|
|
||
| private def normalize(name: String, isCaseSensitive: Boolean): String = | ||
| if (isCaseSensitive) { | ||
| name | ||
| } else { | ||
| name.toLowerCase(Locale.ROOT) | ||
| } | ||
| } | ||
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,103 @@ | ||
| /* | ||
| * 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.kyuubi.spark.connector.hive | ||
|
|
||
| import scala.annotation.tailrec | ||
|
|
||
| import org.apache.spark.sql.{Row, SparkSession} | ||
| import org.apache.spark.sql.connector.read.{Scan, SupportsRuntimeFiltering} | ||
| import org.apache.spark.sql.execution.SparkPlan | ||
| import org.apache.spark.sql.execution.adaptive.AdaptiveSparkPlanExec | ||
| import org.apache.spark.sql.execution.datasources.v2.BatchScanExec | ||
|
|
||
| import org.apache.kyuubi.spark.connector.hive.read.HiveScan | ||
|
|
||
| class DynamicPartitionPruningSuite extends KyuubiHiveTest { | ||
|
|
||
| private def findScan(spark: SparkSession, sql: String, tableNameHint: String): Scan = { | ||
| // Match on `HiveScan.catalogTable` rather than the node's `toString` because | ||
| // `BatchScanExec.toString` shape differs across Spark versions. | ||
| def matchesHint(b: BatchScanExec): Boolean = b.scan match { | ||
| case h: HiveScan => h.catalogTable.identifier.table == tableNameHint | ||
| case _ => false | ||
| } | ||
|
|
||
| @tailrec | ||
| def findBatchScan(plan: SparkPlan): Option[BatchScanExec] = plan match { | ||
| case aqe: AdaptiveSparkPlanExec => findBatchScan(aqe.inputPlan) | ||
| case _ => plan.collectFirst { | ||
| case b: BatchScanExec if matchesHint(b) => b | ||
| } | ||
| } | ||
|
|
||
| val exec = findBatchScan(spark.sql(sql).queryExecution.executedPlan) | ||
| assert(exec.isDefined) | ||
| exec.get.scan | ||
| } | ||
|
|
||
| test("HiveScan supports DPP runtime filtering on partition columns") { | ||
| Seq( | ||
| ("true", Seq("dt")), | ||
| ("false", Seq.empty[String])).foreach { case (enabled, expectedFilterAttrs) => | ||
| withSparkSession(Map( | ||
| "hive.exec.dynamic.partition.mode" -> "nonstrict", | ||
| "spark.sql.kyuubi.hive.connector.read.runtimeFilter.enabled" -> enabled)) { spark => | ||
| val suffix = if (enabled == "true") "on" else "off" | ||
| val fact = s"hive.default.dpp_fact_$suffix" | ||
| val dim = s"hive.default.dpp_dim_$suffix" | ||
|
|
||
| withTable(fact, dim) { | ||
| spark.sql( | ||
| s""" | ||
| | CREATE TABLE $fact (id INT, v STRING) PARTITIONED BY (dt STRING) | ||
| | STORED AS TEXTFILE | ||
| |""".stripMargin).collect() | ||
| spark.sql(s"INSERT INTO $fact PARTITION (dt='2026-01-01') VALUES (1, 'a'), (2, 'b')") | ||
| spark.sql(s"INSERT INTO $fact PARTITION (dt='2026-05-01') VALUES (3, 'c'), (4, 'd')") | ||
| spark.sql(s"INSERT INTO $fact PARTITION (dt='2026-09-01') VALUES (5, 'e'), (6, 'f')") | ||
|
|
||
| spark.sql( | ||
| s""" | ||
| | CREATE TABLE $dim (dt STRING, tag STRING) | ||
| | STORED AS TEXTFILE | ||
| |""".stripMargin).collect() | ||
| spark.sql(s"INSERT INTO $dim VALUES ('2026-05-01', 'target')") | ||
|
|
||
| val sql = | ||
| s""" | ||
| | SELECT f.id, f.v, f.dt | ||
| | FROM $fact f JOIN $dim d ON f.dt = d.dt | ||
| | WHERE d.tag = 'target' | ||
| |""".stripMargin | ||
|
|
||
| checkAnswer( | ||
| spark.sql(sql), | ||
| Seq( | ||
| Row(3, "c", "2026-05-01"), | ||
| Row(4, "d", "2026-05-01"))) | ||
|
|
||
| val scan = findScan(spark, sql, fact.split('.').last) | ||
| assert(scan.isInstanceOf[SupportsRuntimeFiltering]) | ||
| val filterAttrs = scan.asInstanceOf[SupportsRuntimeFiltering] | ||
| .filterAttributes().map(_.fieldNames().mkString(".")) | ||
| assert(filterAttrs.toSeq == expectedFilterAttrs) | ||
| } | ||
| } | ||
| } | ||
| } | ||
| } |
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think in “To restore the legacy behavior, set it to true.” it should be false instead of true?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Oh, my mistake, thanks for review! I've already made the changes.