You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by GitBox <gi...@apache.org> on 2022/05/13 13:50:51 UTC
[GitHub] [spark] cloud-fan opened a new pull request, #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
cloud-fan opened a new pull request, #36541:
URL: https://github.com/apache/spark/pull/36541
<!--
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.
-->
This PR simplifies the planning of limit and offset:
1. Unify the semantics of physical plans that need to deal with limit + offset. These physical plans always do limit first, then offset. The planner rule should set limit and offset properly, for different plans, such as limit + offset and offset + limit.
2. Do not duplicate the planning of `TakeOrderedAndProjectExec`. There is no difference if the `TakeOrderedAndProjectExec` is the root node or not. We don't need to handle it specially under `ReturnAnswer`.
3. Let `GlobalLimitExec` to handle offset as well, so that we can remove `GlobalLimitAndOffsetExec`. This matches `CollectLimitExec`.
### 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.
4. If you fix a bug, you can clarify why it is a bug.
-->
### 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'.
-->
### 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.
-->
--
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] cloud-fan commented on pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on PR #36541:
URL: https://github.com/apache/spark/pull/36541#issuecomment-1127922415
thanks for the review, merging to master!
--
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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r872964108
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -82,52 +82,45 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
- case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child)))
Review Comment:
missing too ?
--
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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873410416
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/limit.scala:
##########
@@ -278,9 +265,9 @@ case class GlobalLimitAndOffsetExec(
}
/**
- * Take the first limit elements as defined by the sortOrder, and do projection if needed.
- * This is logically equivalent to having a Limit operator after a [[SortExec]] operator,
- * or having a [[ProjectExec]] operator between them.
+ * Take the first `limit`` elements as defined by the sortOrder, then drop the first `offset`
Review Comment:
```suggestion
* Take the first `limit` elements as defined by the sortOrder, then drop the first `offset`
```
--
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] cloud-fan commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873317698
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -82,52 +82,45 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
Review Comment:
As I mentioned in the PR description, we don't need to plan `TakeOrderedAndProjectExec` under `ReturnAnswer`, as we don't have special logic for it. It will still be planned in the normal code path, which is `case other => planLater(other) :: Nil` and we do have planner rule to match `Limit(IntegerLiteral(limit), Sort(order, true, child))`
--
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] cloud-fan commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873386778
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -82,52 +82,45 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
Review Comment:
handled by `planTakeOrdered`
--
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] cloud-fan commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873423506
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
+ Some((offset, globalLimit, child))
Review Comment:
This pattern match is to match a logical offset + limit, but we care more about semantics here. Returning `localLimit` is super confusing.
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
+ Some((offset, globalLimit, child))
Review Comment:
This pattern match is to match a logical offset + limit, and we care more about semantics here. Returning `localLimit` is super confusing.
--
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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873425906
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
+ Some((offset, globalLimit, child))
Review Comment:
OK
--
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] cloud-fan commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873317698
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -82,52 +82,45 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
Review Comment:
As I mentioned in the PR description, we don't need to plan `TakeOrderedAndProjectExec` under `ReturnAnswer`, as we don't have special logic for it. It will still be planned in the normal code path, which is `case other => planLater(other) :: Nil` and we do have planner rule to match `Limit(IntegerLiteral(limit), Sort(order, true, child))`
--
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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873406676
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -81,55 +81,56 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
*/
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
- case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
- case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child)))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
+ // Call `planTakeOrdered` first which matches a larger plan.
+ case ReturnAnswer(rootPlan) => planTakeOrdered(rootPlan).getOrElse(rootPlan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, child) =>
+ CollectLimitExec(limit = limit, child = planLater(child), offset = offset)
+ case OffsetAndLimit(offset, limit, child) =>
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ CollectLimitExec(limit = offset + limit, child = planLater(child), offset = offset)
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: 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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873409826
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/limit.scala:
##########
@@ -215,61 +211,52 @@ case class LocalLimitExec(limit: Int, child: SparkPlan) extends BaseLimitExec {
}
/**
- * Take the first `limit` elements of the child's single output partition.
- */
-case class GlobalLimitExec(limit: Int, child: SparkPlan) extends BaseLimitExec {
-
- override def requiredChildDistribution: List[Distribution] = AllTuples :: Nil
-
- override protected def withNewChildInternal(newChild: SparkPlan): SparkPlan =
- copy(child = newChild)
-}
-
-/**
- * Skip the first `offset` elements then take the first `limit` of the following elements in
- * the child's single output partition.
+ * Take the first `limit` elements and then drop the first `offset` elements in the child's single
+ * output partition.
*/
-case class GlobalLimitAndOffsetExec(
- limit: Int = -1,
- offset: Int,
- child: SparkPlan) extends BaseLimitExec {
- assert(offset > 0)
+case class GlobalLimitExec(limit: Int = -1, child: SparkPlan, offset: Int = 0)
+ extends BaseLimitExec {
+ assert(limit >= 0 || (limit == -1 && offset > 0))
override def requiredChildDistribution: List[Distribution] = AllTuples :: Nil
- override def doExecute(): RDD[InternalRow] = if (limit >= 0) {
- child.execute().mapPartitionsInternal(iter => iter.slice(offset, limit + offset))
- } else {
- child.execute().mapPartitionsInternal(iter => iter.drop(offset))
+ override def doExecute(): RDD[InternalRow] = {
+ if (offset > 0) {
+ if (limit >= 0) {
+ child.execute().mapPartitionsInternal(iter => iter.slice(offset, limit))
+ } else {
+ child.execute().mapPartitionsInternal(iter => iter.drop(offset))
+ }
+ } else {
+ super.doExecute()
+ }
}
- private lazy val skipTerm = BaseLimitExec.newLimitCountTerm()
-
override def doConsume(ctx: CodegenContext, input: Seq[ExprCode], row: ExprCode): String = {
- ctx.addMutableState(
- CodeGenerator.JAVA_INT, skipTerm, forceInline = true, useFreshName = false)
- if (limit >= 0) {
- // The counter name is already obtained by the upstream operators via `limitNotReachedChecks`.
- // Here we have to inline it to not change its name. This is fine as we won't have many limit
- // operators in one query.
- ctx.addMutableState(
- CodeGenerator.JAVA_INT, countTerm, forceInline = true, useFreshName = false)
- s"""
- | if ($skipTerm < $offset) {
- | $skipTerm += 1;
- | } else if ($countTerm < $limit) {
- | $countTerm += 1;
- | ${consume(ctx, input)}
- | }
+ if (offset > 0) {
+ val skipTerm = ctx.addMutableState(CodeGenerator.JAVA_INT, "rowsSkipped", forceInline = true)
+ if (limit > 0) {
+ // In codegen, we skip the first `offset` rows, then take the first `limit - offset` rows.
+ val finalLimit = limit - offset
+ s"""
+ | if ($skipTerm < $offset) {
+ | $skipTerm += 1;
+ | } else if ($countTerm < $finalLimit) {
+ | $countTerm += 1;
+ | ${consume(ctx, input)}
+ | }
""".stripMargin
- } else {
- s"""
- | if ($skipTerm < $offset) {
- | $skipTerm += 1;
- | } else {
- | ${consume(ctx, input)}
- | }
+ } else {
+ s"""
+ | if ($skipTerm < $offset) {
+ | $skipTerm += 1;
+ | } else {
+ | ${consume(ctx, input)}
+ | }
""".stripMargin
Review Comment:
```suggestion
""".stripMargin
```
--
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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873402602
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -81,55 +81,56 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
*/
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
- case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
- case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child)))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
+ // Call `planTakeOrdered` first which matches a larger plan.
+ case ReturnAnswer(rootPlan) => planTakeOrdered(rootPlan).getOrElse(rootPlan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, child) =>
+ CollectLimitExec(limit = limit, child = planLater(child), offset = offset)
+ case OffsetAndLimit(offset, limit, child) =>
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ CollectLimitExec(limit = offset + limit, child = planLater(child), offset = offset)
case Limit(IntegerLiteral(limit), child) =>
- CollectLimitExec(limit, planLater(child)) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Sort(order, true, child)) if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, child.output, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Project(projectList, Sort(order, true, child)))
- if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, projectList, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset), child) =>
- CollectLimitExec(limit, planLater(child), offset) :: Nil
+ CollectLimitExec(limit = limit, child = planLater(child))
case logical.Offset(IntegerLiteral(offset), child) =>
- CollectLimitExec(child = planLater(child), offset = offset) :: Nil
+ CollectLimitExec(child = planLater(child), offset = offset)
case Tail(IntegerLiteral(limit), child) =>
- CollectTailExec(limit, planLater(child)) :: Nil
- case other => planLater(other) :: Nil
- }
+ CollectTailExec(limit, planLater(child))
+ case other => planLater(other)
+ }) :: Nil
+
+ case other => planTakeOrdered(other).toSeq
+ }
+
+ private def planTakeOrdered(plan: LogicalPlan): Option[SparkPlan] = plan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, Sort(order, true, child))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, child.output, planLater(child), offset))
+ case LimitAndOffset(limit, offset, Project(projectList, Sort(order, true, child)))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, projectList, planLater(child), offset))
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ case OffsetAndLimit(offset, limit, Sort(order, true, child))
+ if offset + limit < conf.topKSortFallbackThreshold =>
Review Comment:
If we use `localLimit` directly, we can avoid `offset + limit` here.
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -81,55 +81,56 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
*/
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
- case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
- case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child)))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
+ // Call `planTakeOrdered` first which matches a larger plan.
+ case ReturnAnswer(rootPlan) => planTakeOrdered(rootPlan).getOrElse(rootPlan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, child) =>
+ CollectLimitExec(limit = limit, child = planLater(child), offset = offset)
+ case OffsetAndLimit(offset, limit, child) =>
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ CollectLimitExec(limit = offset + limit, child = planLater(child), offset = offset)
case Limit(IntegerLiteral(limit), child) =>
- CollectLimitExec(limit, planLater(child)) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Sort(order, true, child)) if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, child.output, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Project(projectList, Sort(order, true, child)))
- if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, projectList, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset), child) =>
- CollectLimitExec(limit, planLater(child), offset) :: Nil
+ CollectLimitExec(limit = limit, child = planLater(child))
case logical.Offset(IntegerLiteral(offset), child) =>
- CollectLimitExec(child = planLater(child), offset = offset) :: Nil
+ CollectLimitExec(child = planLater(child), offset = offset)
case Tail(IntegerLiteral(limit), child) =>
- CollectTailExec(limit, planLater(child)) :: Nil
- case other => planLater(other) :: Nil
- }
+ CollectTailExec(limit, planLater(child))
+ case other => planLater(other)
+ }) :: Nil
+
+ case other => planTakeOrdered(other).toSeq
+ }
+
+ private def planTakeOrdered(plan: LogicalPlan): Option[SparkPlan] = plan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, Sort(order, true, child))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, child.output, planLater(child), offset))
+ case LimitAndOffset(limit, offset, Project(projectList, Sort(order, true, child)))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, projectList, planLater(child), offset))
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ case OffsetAndLimit(offset, limit, Sort(order, true, child))
+ if offset + limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ offset + limit, order, child.output, planLater(child), offset))
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: 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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r872964095
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -82,52 +82,45 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
Review Comment:
It seems we missing this case.
--
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] cloud-fan commented on pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on PR #36541:
URL: https://github.com/apache/spark/pull/36541#issuecomment-1126081836
@beliefer @dtenedor
--
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] cloud-fan commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873425084
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
+ Some((offset, globalLimit, child))
Review Comment:
It's better to go for better readability, instead of saving a bit typing.
--
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] dtenedor commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
dtenedor commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r872642301
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
Review Comment:
```suggestion
if globalLimit + offset == localLimit =>
```
--
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] cloud-fan closed pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
cloud-fan closed pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
URL: https://github.com/apache/spark/pull/36541
--
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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873403382
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -814,12 +815,19 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
case logical.LocalRelation(output, data, _) =>
LocalTableScanExec(output, data) :: Nil
case CommandResult(output, _, plan, data) => CommandResultExec(output, plan, data) :: Nil
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, child) =>
+ GlobalLimitExec(limit, planLater(child), offset) :: Nil
+ case OffsetAndLimit(offset, limit, child) =>
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ GlobalLimitExec(limit = offset + limit, child = planLater(child), offset = offset) :: Nil
Review Comment:
ditto
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -81,55 +81,56 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
*/
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
- case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
- case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child)))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
+ // Call `planTakeOrdered` first which matches a larger plan.
+ case ReturnAnswer(rootPlan) => planTakeOrdered(rootPlan).getOrElse(rootPlan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, child) =>
+ CollectLimitExec(limit = limit, child = planLater(child), offset = offset)
+ case OffsetAndLimit(offset, limit, child) =>
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ CollectLimitExec(limit = offset + limit, child = planLater(child), offset = offset)
case Limit(IntegerLiteral(limit), child) =>
- CollectLimitExec(limit, planLater(child)) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Sort(order, true, child)) if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, child.output, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Project(projectList, Sort(order, true, child)))
- if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, projectList, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset), child) =>
- CollectLimitExec(limit, planLater(child), offset) :: Nil
+ CollectLimitExec(limit = limit, child = planLater(child))
case logical.Offset(IntegerLiteral(offset), child) =>
- CollectLimitExec(child = planLater(child), offset = offset) :: Nil
+ CollectLimitExec(child = planLater(child), offset = offset)
case Tail(IntegerLiteral(limit), child) =>
- CollectTailExec(limit, planLater(child)) :: Nil
- case other => planLater(other) :: Nil
- }
+ CollectTailExec(limit, planLater(child))
+ case other => planLater(other)
+ }) :: Nil
+
+ case other => planTakeOrdered(other).toSeq
+ }
+
+ private def planTakeOrdered(plan: LogicalPlan): Option[SparkPlan] = plan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, Sort(order, true, child))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, child.output, planLater(child), offset))
+ case LimitAndOffset(limit, offset, Project(projectList, Sort(order, true, child)))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, projectList, planLater(child), offset))
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ case OffsetAndLimit(offset, limit, Sort(order, true, child))
+ if offset + limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ offset + limit, order, child.output, planLater(child), offset))
+ case OffsetAndLimit(offset, limit, Project(projectList, Sort(order, true, child)))
+ if offset + limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ offset + limit, order, projectList, planLater(child), offset))
Review Comment:
ditto
##########
sql/core/src/main/scala/org/apache/spark/sql/execution/SparkStrategies.scala:
##########
@@ -81,55 +81,56 @@ abstract class SparkStrategies extends QueryPlanner[SparkPlan] {
*/
object SpecialLimits extends Strategy {
override def apply(plan: LogicalPlan): Seq[SparkPlan] = plan match {
- case ReturnAnswer(rootPlan) => rootPlan match {
- case Limit(IntegerLiteral(limit), Sort(order, true, child))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, child.output, planLater(child)) :: Nil
- case Limit(IntegerLiteral(limit), Project(projectList, Sort(order, true, child)))
- if limit < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(limit, order, projectList, planLater(child)) :: Nil
+ // Call `planTakeOrdered` first which matches a larger plan.
+ case ReturnAnswer(rootPlan) => planTakeOrdered(rootPlan).getOrElse(rootPlan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, child) =>
+ CollectLimitExec(limit = limit, child = planLater(child), offset = offset)
+ case OffsetAndLimit(offset, limit, child) =>
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ CollectLimitExec(limit = offset + limit, child = planLater(child), offset = offset)
case Limit(IntegerLiteral(limit), child) =>
- CollectLimitExec(limit, planLater(child)) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Sort(order, true, child)) if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, child.output, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset),
- Project(projectList, Sort(order, true, child)))
- if limit + offset < conf.topKSortFallbackThreshold =>
- TakeOrderedAndProjectExec(
- limit, order, projectList, planLater(child), offset) :: Nil
- case LimitAndOffset(IntegerLiteral(limit), IntegerLiteral(offset), child) =>
- CollectLimitExec(limit, planLater(child), offset) :: Nil
+ CollectLimitExec(limit = limit, child = planLater(child))
case logical.Offset(IntegerLiteral(offset), child) =>
- CollectLimitExec(child = planLater(child), offset = offset) :: Nil
+ CollectLimitExec(child = planLater(child), offset = offset)
case Tail(IntegerLiteral(limit), child) =>
- CollectTailExec(limit, planLater(child)) :: Nil
- case other => planLater(other) :: Nil
- }
+ CollectTailExec(limit, planLater(child))
+ case other => planLater(other)
+ }) :: Nil
+
+ case other => planTakeOrdered(other).toSeq
+ }
+
+ private def planTakeOrdered(plan: LogicalPlan): Option[SparkPlan] = plan match {
+ // We should match the combination of limit and offset first, to get the optimal physical
+ // plan, instead of planning limit and offset separately.
+ case LimitAndOffset(limit, offset, Sort(order, true, child))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, child.output, planLater(child), offset))
+ case LimitAndOffset(limit, offset, Project(projectList, Sort(order, true, child)))
+ if limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ limit, order, projectList, planLater(child), offset))
+ // 'Offset a' then 'Limit b' is the same as 'Limit a + b' then 'Offset a'.
+ case OffsetAndLimit(offset, limit, Sort(order, true, child))
+ if offset + limit < conf.topKSortFallbackThreshold =>
+ Some(TakeOrderedAndProjectExec(
+ offset + limit, order, child.output, planLater(child), offset))
+ case OffsetAndLimit(offset, limit, Project(projectList, Sort(order, true, child)))
+ if offset + limit < conf.topKSortFallbackThreshold =>
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: 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] beliefer commented on a diff in pull request #36541: [SPARK-39180][SQL] Simplify the planning of limit and offset
Posted by GitBox <gi...@apache.org>.
beliefer commented on code in PR #36541:
URL: https://github.com/apache/spark/pull/36541#discussion_r873401681
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
+ Some((offset, globalLimit, child))
Review Comment:
In fact, we not use globalLimit in physical plan.
It seems we can return localLimit here. Then we can avoid `+` in physical plan.
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/plans/logical/basicLogicalOperators.scala:
##########
@@ -1303,12 +1303,25 @@ case class LocalLimit(limitExpr: Expression, child: LogicalPlan) extends OrderPr
copy(child = newChild)
}
+object OffsetAndLimit {
+ def unapply(p: GlobalLimit): Option[(Int, Int, LogicalPlan)] = {
+ p match {
+ // Optimizer pushes local limit through offset, so we need to match the plan this way.
+ case GlobalLimit(IntegerLiteral(globalLimit),
+ Offset(IntegerLiteral(offset),
+ LocalLimit(IntegerLiteral(localLimit), child)))
+ if globalLimit + offset == localLimit =>
+ Some((offset, globalLimit, child))
Review Comment:
In fact, we not use `globalLimit` in physical plan.
It seems we can return `localLimit` here. Then we can avoid `+` in physical plan.
--
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