You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "vinodkc (via GitHub)" <gi...@apache.org> on 2023/07/15 00:56:38 UTC
[GitHub] [spark] vinodkc opened a new pull request, #42014: [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec
vinodkc opened a new pull request, #42014:
URL: https://github.com/apache/spark/pull/42014
<!--
Thanks for sending a pull request! Here are some tips for you:
1. If this is your first time, please read our contributor guidelines: https://spark.apache.org/contributing.html
2. Ensure you have added or run the appropriate tests for your PR: https://spark.apache.org/developer-tools.html
3. If the PR is unfinished, add '[WIP]' in your PR title, e.g., '[WIP][SPARK-XXXX] Your PR title ...'.
4. Be sure to keep the PR description updated to reflect all changes.
5. Please write your PR title to summarize what this PR proposes.
6. If possible, provide a concise example to reproduce the issue for a faster review.
7. If you want to add a new configuration, please read the guideline first for naming configurations in
'core/src/main/scala/org/apache/spark/internal/config/ConfigEntry.scala'.
8. If you want to add or modify an error type or message, please read the guideline first in
'core/src/main/resources/error/README.md'.
-->
### What changes were proposed in this pull request?
<!--
Please clarify what changes you are proposing. The purpose of this section is to outline the changes and how this PR fixes the issue.
If possible, please consider writing useful notes for better and faster reviews in your PR. See the examples below.
1. If you refactor some codes with changing classes, showing the class hierarchy will help reviewers.
2. If you fix some SQL features, you can provide some references of other DBMSes.
3. If there is design documentation, please add the link.
4. If there is a discussion in the mailing list, please add the link.
-->
SQL operators `ArrowEvalPythonUDTFExec` & `BatchEvalPythonUDTFExec` are updated to use the `PartitionEvaluator` API to do execution.
### Why are the changes needed?
<!--
Please clarify why the changes are needed. For instance,
1. If you propose a new API, clarify the use case for a new API.
2. If you fix a bug, you can clarify why it is a bug.
-->
To avoid the use of lambda during distributed execution.
Ref: SPARK-43061 for more details.
### Does this PR introduce _any_ user-facing change?
<!--
Note that it means *any* user-facing change including all aspects such as the documentation fix.
If yes, please clarify the previous behavior and the change this PR proposes - provide the console output, description and/or an example to show the behavior difference if possible.
If possible, please also clarify if this is a user-facing change compared to the released Spark versions or within the unreleased branches such as master.
If no, write 'No'.
-->
No
### How was this patch tested?
<!--
If tests were added, say they were added here. Please make sure to add some test cases that check the changes thoroughly including negative and positive cases if possible.
If it was tested in a way different from regular unit tests, please clarify how you tested step by step, ideally copy and paste-able, so that other reviewers can test and check, and descendants can verify in the future.
If tests were not added, please describe why they were not added and/or why it was difficult to add.
If benchmark tests were added, please run the benchmarks in GitHub Actions for the consistent environment, and the instructions could accord to: https://spark.apache.org/developer-tools.html#github-workflow-benchmarks.
-->
Existing test cases. Once all SQL operators are refactored, will enable `spark.sql.execution.usePartitionEvaluator` by default, so all tests cover this code path.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] Hisoka-X commented on a diff in pull request #42014: [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec
Posted by "Hisoka-X (via GitHub)" <gi...@apache.org>.
Hisoka-X commented on code in PR #42014:
URL: https://github.com/apache/spark/pull/42014#discussion_r1264316395
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/python/EvalPythonUDTFExec.scala:
##########
@@ -40,85 +33,19 @@ trait EvalPythonUDTFExec extends UnaryExecNode {
def resultAttrs: Seq[Attribute]
+ protected def evaluatorFactory: EvalPythonUDTFEvaluatorFactory
+
override def output: Seq[Attribute] = requiredChildOutput ++ resultAttrs
override def producedAttributes: AttributeSet = AttributeSet(resultAttrs)
- protected def evaluate(
- argOffsets: Array[Int],
- iter: Iterator[InternalRow],
- schema: StructType,
- context: TaskContext): Iterator[Iterator[InternalRow]]
-
protected override def doExecute(): RDD[InternalRow] = {
val inputRDD = child.execute().map(_.copy())
-
- inputRDD.mapPartitions { iter =>
- val context = TaskContext.get()
- val contextAwareIterator = new ContextAwareIterator(context, iter)
-
- // The queue used to buffer input rows so we can drain it to
- // combine input with output from Python.
- val queue = HybridRowQueue(context.taskMemoryManager(),
- new File(Utils.getLocalDir(SparkEnv.get.conf)), child.output.length)
- context.addTaskCompletionListener[Unit] { ctx =>
- queue.close()
- }
-
- // flatten all the arguments
- val allInputs = new ArrayBuffer[Expression]
- val dataTypes = new ArrayBuffer[DataType]
- val argOffsets = udtf.children.map { e =>
- if (allInputs.exists(_.semanticEquals(e))) {
- allInputs.indexWhere(_.semanticEquals(e))
- } else {
- allInputs += e
- dataTypes += e.dataType
- allInputs.length - 1
- }
- }.toArray
- val projection = MutableProjection.create(allInputs.toSeq, child.output)
- projection.initialize(context.partitionId())
- val schema = StructType(dataTypes.zipWithIndex.map { case (dt, i) =>
- StructField(s"_$i", dt)
- }.toArray)
-
- // Add rows to the queue to join later with the result.
- // Also keep track of the number rows added to the queue.
- // This is needed to process extra output rows from the `terminate()` call of the UDTF.
- var count = 0L
- val projectedRowIter = contextAwareIterator.map { inputRow =>
- queue.add(inputRow.asInstanceOf[UnsafeRow])
- count += 1
- projection(inputRow)
- }
-
- val outputRowIterator = evaluate(argOffsets, projectedRowIter, schema, context)
-
- val pruneChildForResult: InternalRow => InternalRow =
- if (child.outputSet == AttributeSet(requiredChildOutput)) {
- identity
- } else {
- UnsafeProjection.create(requiredChildOutput, child.output)
- }
-
- val joined = new JoinedRow
- val resultProj = UnsafeProjection.create(output, output)
-
- outputRowIterator.flatMap { outputRows =>
- // If `count` is greater than zero, it means there are remaining input rows in the queue.
- // In this case, the output rows of the UDTF are joined with the corresponding input row
- // in the queue.
- if (count > 0) {
- val left = queue.remove()
- count -= 1
- joined.withLeft(pruneChildForResult(left))
- }
- // If `count` is zero, it means all input rows have been consumed. Any additional rows
- // from the UDTF are from the `terminate()` call. We leave the left side as the last
- // element of its child output to keep it consistent with the Generate implementation
- // and Hive UDTFs.
- outputRows.map(r => resultProj(joined.withRight(r)))
+ if (conf.usePartitionEvaluator) {
+ inputRDD.mapPartitionsWithEvaluator(evaluatorFactory)
+ } else {
+ inputRDD.mapPartitions { iter =>
+ evaluatorFactory.createEvaluator().eval(0, iter)
Review Comment:
Hardcoding `partIndex` may lead to unpredictable situations in maintaining `EvalPythonUDTFEvaluatorFactory`, if subsequent `EvalPythonUDTFEvaluatorFactory` needs to use `partIndex`.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] vinodkc commented on a diff in pull request #42014: [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #42014:
URL: https://github.com/apache/spark/pull/42014#discussion_r1264555511
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/python/EvalPythonUDTFExec.scala:
##########
@@ -40,85 +33,19 @@ trait EvalPythonUDTFExec extends UnaryExecNode {
def resultAttrs: Seq[Attribute]
+ protected def evaluatorFactory: EvalPythonUDTFEvaluatorFactory
+
override def output: Seq[Attribute] = requiredChildOutput ++ resultAttrs
override def producedAttributes: AttributeSet = AttributeSet(resultAttrs)
- protected def evaluate(
- argOffsets: Array[Int],
- iter: Iterator[InternalRow],
- schema: StructType,
- context: TaskContext): Iterator[Iterator[InternalRow]]
-
protected override def doExecute(): RDD[InternalRow] = {
val inputRDD = child.execute().map(_.copy())
-
- inputRDD.mapPartitions { iter =>
- val context = TaskContext.get()
- val contextAwareIterator = new ContextAwareIterator(context, iter)
-
- // The queue used to buffer input rows so we can drain it to
- // combine input with output from Python.
- val queue = HybridRowQueue(context.taskMemoryManager(),
- new File(Utils.getLocalDir(SparkEnv.get.conf)), child.output.length)
- context.addTaskCompletionListener[Unit] { ctx =>
- queue.close()
- }
-
- // flatten all the arguments
- val allInputs = new ArrayBuffer[Expression]
- val dataTypes = new ArrayBuffer[DataType]
- val argOffsets = udtf.children.map { e =>
- if (allInputs.exists(_.semanticEquals(e))) {
- allInputs.indexWhere(_.semanticEquals(e))
- } else {
- allInputs += e
- dataTypes += e.dataType
- allInputs.length - 1
- }
- }.toArray
- val projection = MutableProjection.create(allInputs.toSeq, child.output)
- projection.initialize(context.partitionId())
- val schema = StructType(dataTypes.zipWithIndex.map { case (dt, i) =>
- StructField(s"_$i", dt)
- }.toArray)
-
- // Add rows to the queue to join later with the result.
- // Also keep track of the number rows added to the queue.
- // This is needed to process extra output rows from the `terminate()` call of the UDTF.
- var count = 0L
- val projectedRowIter = contextAwareIterator.map { inputRow =>
- queue.add(inputRow.asInstanceOf[UnsafeRow])
- count += 1
- projection(inputRow)
- }
-
- val outputRowIterator = evaluate(argOffsets, projectedRowIter, schema, context)
-
- val pruneChildForResult: InternalRow => InternalRow =
- if (child.outputSet == AttributeSet(requiredChildOutput)) {
- identity
- } else {
- UnsafeProjection.create(requiredChildOutput, child.output)
- }
-
- val joined = new JoinedRow
- val resultProj = UnsafeProjection.create(output, output)
-
- outputRowIterator.flatMap { outputRows =>
- // If `count` is greater than zero, it means there are remaining input rows in the queue.
- // In this case, the output rows of the UDTF are joined with the corresponding input row
- // in the queue.
- if (count > 0) {
- val left = queue.remove()
- count -= 1
- joined.withLeft(pruneChildForResult(left))
- }
- // If `count` is zero, it means all input rows have been consumed. Any additional rows
- // from the UDTF are from the `terminate()` call. We leave the left side as the last
- // element of its child output to keep it consistent with the Generate implementation
- // and Hive UDTFs.
- outputRows.map(r => resultProj(joined.withRight(r)))
+ if (conf.usePartitionEvaluator) {
+ inputRDD.mapPartitionsWithEvaluator(evaluatorFactory)
+ } else {
+ inputRDD.mapPartitions { iter =>
+ evaluatorFactory.createEvaluator().eval(0, iter)
Review Comment:
Done. Used mapPartitionsWithIndexInternal
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec [spark]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] commented on PR #42014:
URL: https://github.com/apache/spark/pull/42014#issuecomment-1793589610
We're closing this PR because it hasn't been updated in a while. This isn't a judgement on the merit of the PR in any way. It's just a way of keeping the PR queue manageable.
If you'd like to revive this PR, please reopen it and ask a committer to remove the Stale tag!
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
[GitHub] [spark] vinodkc commented on a diff in pull request #42014: [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec
Posted by "vinodkc (via GitHub)" <gi...@apache.org>.
vinodkc commented on code in PR #42014:
URL: https://github.com/apache/spark/pull/42014#discussion_r1264555511
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/python/EvalPythonUDTFExec.scala:
##########
@@ -40,85 +33,19 @@ trait EvalPythonUDTFExec extends UnaryExecNode {
def resultAttrs: Seq[Attribute]
+ protected def evaluatorFactory: EvalPythonUDTFEvaluatorFactory
+
override def output: Seq[Attribute] = requiredChildOutput ++ resultAttrs
override def producedAttributes: AttributeSet = AttributeSet(resultAttrs)
- protected def evaluate(
- argOffsets: Array[Int],
- iter: Iterator[InternalRow],
- schema: StructType,
- context: TaskContext): Iterator[Iterator[InternalRow]]
-
protected override def doExecute(): RDD[InternalRow] = {
val inputRDD = child.execute().map(_.copy())
-
- inputRDD.mapPartitions { iter =>
- val context = TaskContext.get()
- val contextAwareIterator = new ContextAwareIterator(context, iter)
-
- // The queue used to buffer input rows so we can drain it to
- // combine input with output from Python.
- val queue = HybridRowQueue(context.taskMemoryManager(),
- new File(Utils.getLocalDir(SparkEnv.get.conf)), child.output.length)
- context.addTaskCompletionListener[Unit] { ctx =>
- queue.close()
- }
-
- // flatten all the arguments
- val allInputs = new ArrayBuffer[Expression]
- val dataTypes = new ArrayBuffer[DataType]
- val argOffsets = udtf.children.map { e =>
- if (allInputs.exists(_.semanticEquals(e))) {
- allInputs.indexWhere(_.semanticEquals(e))
- } else {
- allInputs += e
- dataTypes += e.dataType
- allInputs.length - 1
- }
- }.toArray
- val projection = MutableProjection.create(allInputs.toSeq, child.output)
- projection.initialize(context.partitionId())
- val schema = StructType(dataTypes.zipWithIndex.map { case (dt, i) =>
- StructField(s"_$i", dt)
- }.toArray)
-
- // Add rows to the queue to join later with the result.
- // Also keep track of the number rows added to the queue.
- // This is needed to process extra output rows from the `terminate()` call of the UDTF.
- var count = 0L
- val projectedRowIter = contextAwareIterator.map { inputRow =>
- queue.add(inputRow.asInstanceOf[UnsafeRow])
- count += 1
- projection(inputRow)
- }
-
- val outputRowIterator = evaluate(argOffsets, projectedRowIter, schema, context)
-
- val pruneChildForResult: InternalRow => InternalRow =
- if (child.outputSet == AttributeSet(requiredChildOutput)) {
- identity
- } else {
- UnsafeProjection.create(requiredChildOutput, child.output)
- }
-
- val joined = new JoinedRow
- val resultProj = UnsafeProjection.create(output, output)
-
- outputRowIterator.flatMap { outputRows =>
- // If `count` is greater than zero, it means there are remaining input rows in the queue.
- // In this case, the output rows of the UDTF are joined with the corresponding input row
- // in the queue.
- if (count > 0) {
- val left = queue.remove()
- count -= 1
- joined.withLeft(pruneChildForResult(left))
- }
- // If `count` is zero, it means all input rows have been consumed. Any additional rows
- // from the UDTF are from the `terminate()` call. We leave the left side as the last
- // element of its child output to keep it consistent with the Generate implementation
- // and Hive UDTFs.
- outputRows.map(r => resultProj(joined.withRight(r)))
+ if (conf.usePartitionEvaluator) {
+ inputRDD.mapPartitionsWithEvaluator(evaluatorFactory)
+ } else {
+ inputRDD.mapPartitions { iter =>
+ evaluatorFactory.createEvaluator().eval(0, iter)
Review Comment:
Yes, one of the options is to move `doExecute` implementation to concrete classes.
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org
Re: [PR] [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec [spark]
Posted by "github-actions[bot] (via GitHub)" <gi...@apache.org>.
github-actions[bot] closed pull request #42014: [SPARK-44412][SQL] Use PartitionEvaluator API in ArrowEvalPythonUDTFExec & BatchEvalPythonUDTFExec
URL: https://github.com/apache/spark/pull/42014
--
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: reviews-unsubscribe@spark.apache.org
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org