This is an automated email from the ASF dual-hosted git repository.
hongze pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/incubator-gluten.git
The following commit(s) were added to refs/heads/main by this push:
new 16f68a2203 [GLUTEN-7267][CORE][CH] Move schema pruning optimization of
HiveTableScan to an individual post-transform rule (#8008)
16f68a2203 is described below
commit 16f68a2203e70075554e83adf15aec7676ad1d1f
Author: Hongze Zhang <[email protected]>
AuthorDate: Thu Nov 21 09:32:35 2024 +0800
[GLUTEN-7267][CORE][CH] Move schema pruning optimization of HiveTableScan
to an individual post-transform rule (#8008)
---
.../gluten/backendsapi/clickhouse/CHBackend.scala | 4 ---
.../gluten/backendsapi/clickhouse/CHRuleApi.scala | 1 +
.../gluten/backendsapi/BackendSettingsApi.scala | 2 --
.../PruneNestedColumnsInHiveTableScan.scala | 33 ++++++++++++++++++++++
.../columnar/offload/OffloadSingleNodeRules.scala | 11 ++------
.../hive/HiveTableScanNestedColumnPruning.scala | 12 ++++----
6 files changed, 43 insertions(+), 20 deletions(-)
diff --git
a/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHBackend.scala
b/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHBackend.scala
index 966e953199..98d1eabe57 100644
---
a/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHBackend.scala
+++
b/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHBackend.scala
@@ -407,8 +407,4 @@ object CHBackendSettings extends BackendSettingsApi with
Logging {
}
override def supportWindowGroupLimitExec(rankLikeFunction: Expression):
Boolean = true
-
- override def supportHiveTableScanNestedColumnPruning: Boolean =
- GlutenConfig.getConf.enableColumnarHiveTableScanNestedColumnPruning
-
}
diff --git
a/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHRuleApi.scala
b/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHRuleApi.scala
index 3525fc768d..081e815588 100644
---
a/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHRuleApi.scala
+++
b/backends-clickhouse/src/main/scala/org/apache/gluten/backendsapi/clickhouse/CHRuleApi.scala
@@ -98,6 +98,7 @@ object CHRuleApi {
c => intercept(HeuristicTransform.Single(validatorBuilder(c.glutenConf),
rewrites, offloads)))
// Legacy: Post-transform rules.
+ injector.injectPostTransform(_ => PruneNestedColumnsInHiveTableScan)
injector.injectPostTransform(_ => RemoveNativeWriteFilesSortAndProject())
injector.injectPostTransform(c =>
intercept(RewriteTransformer.apply(c.session)))
injector.injectPostTransform(_ => PushDownFilterToScan)
diff --git
a/gluten-substrait/src/main/scala/org/apache/gluten/backendsapi/BackendSettingsApi.scala
b/gluten-substrait/src/main/scala/org/apache/gluten/backendsapi/BackendSettingsApi.scala
index 700571fd28..177d19c0c7 100644
---
a/gluten-substrait/src/main/scala/org/apache/gluten/backendsapi/BackendSettingsApi.scala
+++
b/gluten-substrait/src/main/scala/org/apache/gluten/backendsapi/BackendSettingsApi.scala
@@ -131,6 +131,4 @@ trait BackendSettingsApi {
def supportColumnarArrowUdf(): Boolean = false
def needPreComputeRangeFrameBoundary(): Boolean = false
-
- def supportHiveTableScanNestedColumnPruning(): Boolean = false
}
diff --git
a/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/PruneNestedColumnsInHiveTableScan.scala
b/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/PruneNestedColumnsInHiveTableScan.scala
new file mode 100644
index 0000000000..b8972b6d0e
--- /dev/null
+++
b/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/PruneNestedColumnsInHiveTableScan.scala
@@ -0,0 +1,33 @@
+/*
+ * 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.gluten.extension.columnar
+
+import org.apache.gluten.execution.ProjectExecTransformer
+
+import org.apache.spark.sql.catalyst.rules.Rule
+import org.apache.spark.sql.execution.SparkPlan
+import org.apache.spark.sql.hive.HiveTableScanNestedColumnPruning
+
+// Since https://github.com/apache/incubator-gluten/pull/7268.
+// Used only by CH backend as of now.
+object PruneNestedColumnsInHiveTableScan extends Rule[SparkPlan] {
+ override def apply(plan: SparkPlan): SparkPlan = plan.transformUp {
+ case p: ProjectExecTransformer
+ if HiveTableScanNestedColumnPruning.supportNestedColumnPruning(p) =>
+ HiveTableScanNestedColumnPruning.apply(p)
+ }
+}
diff --git
a/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/offload/OffloadSingleNodeRules.scala
b/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/offload/OffloadSingleNodeRules.scala
index 64448fc60a..0fee61acac 100644
---
a/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/offload/OffloadSingleNodeRules.scala
+++
b/gluten-substrait/src/main/scala/org/apache/gluten/extension/columnar/offload/OffloadSingleNodeRules.scala
@@ -37,7 +37,7 @@ import
org.apache.spark.sql.execution.exchange.{BroadcastExchangeExec, ShuffleEx
import org.apache.spark.sql.execution.joins._
import org.apache.spark.sql.execution.python.{ArrowEvalPythonExec,
BatchEvalPythonExec}
import org.apache.spark.sql.execution.window.{WindowExec,
WindowGroupLimitExecShim}
-import org.apache.spark.sql.hive.{HiveTableScanExecTransformer,
HiveTableScanNestedColumnPruning}
+import org.apache.spark.sql.hive.HiveTableScanExecTransformer
// Exchange transformation.
case class OffloadExchange() extends OffloadSingleNode with LogLevelUtil {
@@ -188,7 +188,7 @@ object OffloadOthers {
// Utility to replace single node within transformed Gluten node.
// Children will be preserved as they are as children of the output node.
//
- // Do not look-up on children on the input node in this rule. Otherwise
+ // Do not look up on children on the input node in this rule. Otherwise
// it may break RAS which would group all the possible input nodes to
// search for validate candidates.
private class ReplaceSingleNode() extends LogLevelUtil with Logging {
@@ -215,11 +215,7 @@ object OffloadOthers {
case plan: ProjectExec =>
val columnarChild = plan.child
logDebug(s"Columnar Processing for ${plan.getClass} is currently
supported.")
- if
(HiveTableScanNestedColumnPruning.supportNestedColumnPruning(plan)) {
- HiveTableScanNestedColumnPruning.apply(plan)
- } else {
- ProjectExecTransformer(plan.projectList, columnarChild)
- }
+ ProjectExecTransformer(plan.projectList, columnarChild)
case plan: HashAggregateExec =>
logDebug(s"Columnar Processing for ${plan.getClass} is currently
supported.")
HashAggregateExecBaseTransformer.from(plan)
@@ -247,7 +243,6 @@ object OffloadOthers {
plan.bucketSpec,
plan.options,
plan.staticPartitions)
-
ColumnarWriteFilesExec(
writeTransformer,
plan.fileFormat,
diff --git
a/gluten-substrait/src/main/scala/org/apache/spark/sql/hive/HiveTableScanNestedColumnPruning.scala
b/gluten-substrait/src/main/scala/org/apache/spark/sql/hive/HiveTableScanNestedColumnPruning.scala
index 7a20e5c37d..70b08a2ed8 100644
---
a/gluten-substrait/src/main/scala/org/apache/spark/sql/hive/HiveTableScanNestedColumnPruning.scala
+++
b/gluten-substrait/src/main/scala/org/apache/spark/sql/hive/HiveTableScanNestedColumnPruning.scala
@@ -16,13 +16,13 @@
*/
package org.apache.spark.sql.hive
-import org.apache.gluten.backendsapi.BackendsApiManager
+import org.apache.gluten.GlutenConfig
import org.apache.gluten.execution.ProjectExecTransformer
import org.apache.spark.internal.Logging
import org.apache.spark.sql.catalyst.catalog.HiveTableRelation
import org.apache.spark.sql.catalyst.expressions._
-import org.apache.spark.sql.execution.{FilterExec, LeafExecNode, ProjectExec,
SparkPlan}
+import org.apache.spark.sql.execution.{FilterExec, LeafExecNode, SparkPlan}
import
org.apache.spark.sql.hive.HiveTableScanExecTransformer.{ORC_INPUT_FORMAT_CLASS,
PARQUET_INPUT_FORMAT_CLASS, TEXT_INPUT_FORMAT_CLASS}
import org.apache.spark.sql.types.{ArrayType, DataType, MapType, StructType}
import org.apache.spark.sql.util.SchemaUtils._
@@ -31,9 +31,9 @@ import org.apache.spark.util.Utils
object HiveTableScanNestedColumnPruning extends Logging {
import org.apache.spark.sql.catalyst.expressions.SchemaPruning._
- def supportNestedColumnPruning(projectExec: ProjectExec): Boolean = {
- if
(BackendsApiManager.getSettings.supportHiveTableScanNestedColumnPruning()) {
- projectExec.child match {
+ def supportNestedColumnPruning(project: ProjectExecTransformer): Boolean = {
+ if (GlutenConfig.getConf.enableColumnarHiveTableScanNestedColumnPruning) {
+ project.child match {
case HiveTableScanExecTransformer(_, relation, _, _) =>
relation.tableMeta.storage.inputFormat match {
case Some(inputFormat)
@@ -60,7 +60,7 @@ object HiveTableScanNestedColumnPruning extends Logging {
def apply(plan: SparkPlan): SparkPlan = {
plan match {
- case ProjectExec(projectList, child) =>
+ case ProjectExecTransformer(projectList, child) =>
child match {
case h: HiveTableScanExecTransformer =>
val newPlan = prunePhysicalColumns(
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
