You are viewing a plain text version of this content. The canonical link for it is here.
Posted to reviews@spark.apache.org by "anchovYu (via GitHub)" <gi...@apache.org> on 2023/08/01 22:06:07 UTC
[GitHub] [spark] anchovYu opened a new pull request, #42276: [WIP] Throw a more tailored error message for having - window - LCA unresolved case
anchovYu opened a new pull request, #42276:
URL: https://github.com/apache/spark/pull/42276
<!--
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.
-->
### 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.
-->
### 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 closed pull request #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan closed pull request #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
URL: https://github.com/apache/spark/pull/42276
--
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 #42276: [WIP] Throw a more tailored error message for having - window - LCA unresolved case
Posted by "Hisoka-X (via GitHub)" <gi...@apache.org>.
Hisoka-X commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1281352199
##########
common/utils/src/main/resources/error/error-classes.json:
##########
@@ -2952,7 +2952,7 @@
},
"LATERAL_COLUMN_ALIAS_IN_AGGREGATE_WITH_WINDOW_AND_HAVING" : {
"message" : [
- "Referencing lateral column alias <lca> in the aggregate query both with window expressions and with having clause. Please rewrite the aggregate query by removing the having clause or removing lateral alias reference in the SELECT list."
+ "Referencing lateral column alias <lca> in the query both with window expressions and with having clause. Please rewrite the aggregate query by removing the having clause or removing lateral alias reference in the SELECT list that has the window expression and retry."
Review Comment:
Please follow https://github.com/apache/spark/blob/ca274fe15cf9f621394759339b6c8d826692d1a1/core/src/test/scala/org/apache/spark/SparkThrowableSuite.scala#L51C1-L56C1 to regenerate doc.
--
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] anchovYu commented on a diff in pull request #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
Posted by "anchovYu (via GitHub)" <gi...@apache.org>.
anchovYu commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1285473731
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,138 +131,166 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written top down recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate)
+ if agg.aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
Review Comment:
Completely copied the original code except adding the final line to recursively iterating children
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,138 +131,166 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written top down recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate)
+ if agg.aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
+ && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ var aliasMap = AttributeMap.empty[AliasEntry]
+ val referencedAliases = collection.mutable.Set.empty[AliasEntry]
+ def unwrapLCAReference(e: NamedExpression): NamedExpression = {
+ e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
+ val aliasEntry = aliasMap.get(lcaRef.a).get
+ // If there is no chaining of lateral column alias reference, push down the alias
+ // and unwrap the LateralColumnAliasReference to the NamedExpression inside
+ // If there is chaining, don't resolve and save to future rounds
+ if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ referencedAliases += aliasEntry
+ lcaRef.ne
+ } else {
+ lcaRef
+ }
+ case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
+ // It shouldn't happen, but restore to unresolved attribute to be safe.
+ UnresolvedAttribute(lcaRef.nameParts)
+ }.asInstanceOf[NamedExpression]
+ }
+ val newProjectList = projectList.zipWithIndex.map {
+ case (a: Alias, idx) =>
+ val lcaResolved = unwrapLCAReference(a)
+ // Insert the original alias instead of rewritten one to detect chained LCA
+ aliasMap += (a.toAttribute -> AliasEntry(a, idx))
+ lcaResolved
+ case (e, _) =>
+ unwrapLCAReference(e)
+ }
- // Check if current Aggregate is eligible to lift up with Project: the aggregate
- // expression only contains: 1) aggregate functions, 2) grouping expressions, 3) leaf
- // expressions excluding attributes not in grouping expressions
- // This check is to prevent unnecessary transformation on invalid plan, to guarantee it
- // throws the same exception. For example, cases like non-aggregate expressions not
- // in group by, once transformed, will throw a different exception: missing input.
- def eligibleToLiftUp(exp: Expression): Boolean = {
- exp match {
- case _: AggregateExpression => true
- case e if groupingExpressions.exists(_.semanticEquals(e)) => true
- case a: Attribute => false
- case s: ScalarSubquery if s.children.nonEmpty
- && !groupingExpressions.exists(_.semanticEquals(s)) => false
- // Manually skip detection on function itself because it can be an aggregate function.
- // This is to avoid expressions like sum(salary) over () eligible to lift up.
- case WindowExpression(function, spec) =>
- function.children.forall(eligibleToLiftUp) && eligibleToLiftUp(spec)
- case e => e.children.forall(eligibleToLiftUp)
- }
+ val newPlan = if (referencedAliases.isEmpty) {
+ p
+ } else {
+ val outerProjectList = collection.mutable.Seq(newProjectList: _*)
+ val innerProjectList =
+ collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
+ referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
+ outerProjectList.update(idx, alias.toAttribute)
+ innerProjectList += alias
}
- if (!aggregateExpressions.forall(eligibleToLiftUp)) {
- return agg
+ p.copy(
+ projectList = outerProjectList.toSeq,
+ child = Project(innerProjectList.toSeq, child)
+ )
+ }
+ newPlan.mapChildren(apply0)
+
+ case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
Review Comment:
Completely copied the original code except adding the final line to recursively applying on the new Project
--
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] anchovYu commented on a diff in pull request #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
Posted by "anchovYu (via GitHub)" <gi...@apache.org>.
anchovYu commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1286512577
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,95 +131,97 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written bottom-up recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 and SPARK-44714 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case pOriginal: Project if ruleApplicableOnOperator(pOriginal, pOriginal.projectList)
+ && pOriginal.projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ val p @ Project(projectList, child) = pOriginal.mapChildren(apply0)
+ var aliasMap = AttributeMap.empty[AliasEntry]
+ val referencedAliases = collection.mutable.Set.empty[AliasEntry]
+ def unwrapLCAReference(e: NamedExpression): NamedExpression = {
+ e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
+ val aliasEntry = aliasMap.get(lcaRef.a).get
+ // If there is no chaining of lateral column alias reference, push down the alias
+ // and unwrap the LateralColumnAliasReference to the NamedExpression inside
+ // If there is chaining, don't resolve and save to future rounds
+ if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ referencedAliases += aliasEntry
+ lcaRef.ne
+ } else {
+ lcaRef
+ }
+ case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
+ // It shouldn't happen, but restore to unresolved attribute to be safe.
+ UnresolvedAttribute(lcaRef.nameParts)
+ }.asInstanceOf[NamedExpression]
+ }
+ val newProjectList = projectList.zipWithIndex.map {
+ case (a: Alias, idx) =>
+ val lcaResolved = unwrapLCAReference(a)
+ // Insert the original alias instead of rewritten one to detect chained LCA
+ aliasMap += (a.toAttribute -> AliasEntry(a, idx))
+ lcaResolved
+ case (e, _) =>
+ unwrapLCAReference(e)
+ }
- // Check if current Aggregate is eligible to lift up with Project: the aggregate
- // expression only contains: 1) aggregate functions, 2) grouping expressions, 3) leaf
- // expressions excluding attributes not in grouping expressions
- // This check is to prevent unnecessary transformation on invalid plan, to guarantee it
- // throws the same exception. For example, cases like non-aggregate expressions not
- // in group by, once transformed, will throw a different exception: missing input.
- def eligibleToLiftUp(exp: Expression): Boolean = {
- exp match {
- case _: AggregateExpression => true
- case e if groupingExpressions.exists(_.semanticEquals(e)) => true
- case a: Attribute => false
- case s: ScalarSubquery if s.children.nonEmpty
- && !groupingExpressions.exists(_.semanticEquals(s)) => false
- // Manually skip detection on function itself because it can be an aggregate function.
- // This is to avoid expressions like sum(salary) over () eligible to lift up.
- case WindowExpression(function, spec) =>
- function.children.forall(eligibleToLiftUp) && eligibleToLiftUp(spec)
- case e => e.children.forall(eligibleToLiftUp)
- }
- }
- if (!aggregateExpressions.forall(eligibleToLiftUp)) {
- return agg
+ if (referencedAliases.isEmpty) {
+ p
+ } else {
+ val outerProjectList = collection.mutable.Seq(newProjectList: _*)
+ val innerProjectList =
+ collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
+ referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
+ outerProjectList.update(idx, alias.toAttribute)
+ innerProjectList += alias
}
+ p.copy(
+ projectList = outerProjectList.toSeq,
+ child = Project(innerProjectList.toSeq, child)
+ )
+ }
+
+ case aggOriginal: Aggregate
+ if ruleApplicableOnOperator(aggOriginal, aggOriginal.aggregateExpressions)
+ && aggOriginal.aggregateExpressions.exists(
+ _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ val agg @ Aggregate(groupingExpressions, aggregateExpressions, _) =
Review Comment:
Bottom up resolution. The rest of code is fully copied and has no change (except the one I commented out).
--
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 #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
Posted by "cloud-fan (via GitHub)" <gi...@apache.org>.
cloud-fan commented on PR #42276:
URL: https://github.com/apache/spark/pull/42276#issuecomment-1669220104
thanks, merging to master/3.5!
--
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] anchovYu commented on a diff in pull request #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
Posted by "anchovYu (via GitHub)" <gi...@apache.org>.
anchovYu commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1286512367
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,95 +131,97 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written bottom-up recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 and SPARK-44714 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case pOriginal: Project if ruleApplicableOnOperator(pOriginal, pOriginal.projectList)
+ && pOriginal.projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ val p @ Project(projectList, child) = pOriginal.mapChildren(apply0)
Review Comment:
Bottom-up resolution. The rest of code is fully copied and has no change.
--
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] anchovYu commented on a diff in pull request #42276: [WIP] Throw a more tailored error message for having - window - LCA unresolved case
Posted by "anchovYu (via GitHub)" <gi...@apache.org>.
anchovYu commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1285474003
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,138 +131,166 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written top down recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate)
+ if agg.aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
+ && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ var aliasMap = AttributeMap.empty[AliasEntry]
+ val referencedAliases = collection.mutable.Set.empty[AliasEntry]
+ def unwrapLCAReference(e: NamedExpression): NamedExpression = {
+ e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
+ val aliasEntry = aliasMap.get(lcaRef.a).get
+ // If there is no chaining of lateral column alias reference, push down the alias
+ // and unwrap the LateralColumnAliasReference to the NamedExpression inside
+ // If there is chaining, don't resolve and save to future rounds
+ if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ referencedAliases += aliasEntry
+ lcaRef.ne
+ } else {
+ lcaRef
+ }
+ case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
+ // It shouldn't happen, but restore to unresolved attribute to be safe.
+ UnresolvedAttribute(lcaRef.nameParts)
+ }.asInstanceOf[NamedExpression]
+ }
+ val newProjectList = projectList.zipWithIndex.map {
+ case (a: Alias, idx) =>
+ val lcaResolved = unwrapLCAReference(a)
+ // Insert the original alias instead of rewritten one to detect chained LCA
+ aliasMap += (a.toAttribute -> AliasEntry(a, idx))
+ lcaResolved
+ case (e, _) =>
+ unwrapLCAReference(e)
+ }
- // Check if current Aggregate is eligible to lift up with Project: the aggregate
- // expression only contains: 1) aggregate functions, 2) grouping expressions, 3) leaf
- // expressions excluding attributes not in grouping expressions
- // This check is to prevent unnecessary transformation on invalid plan, to guarantee it
- // throws the same exception. For example, cases like non-aggregate expressions not
- // in group by, once transformed, will throw a different exception: missing input.
- def eligibleToLiftUp(exp: Expression): Boolean = {
- exp match {
- case _: AggregateExpression => true
- case e if groupingExpressions.exists(_.semanticEquals(e)) => true
- case a: Attribute => false
- case s: ScalarSubquery if s.children.nonEmpty
- && !groupingExpressions.exists(_.semanticEquals(s)) => false
- // Manually skip detection on function itself because it can be an aggregate function.
- // This is to avoid expressions like sum(salary) over () eligible to lift up.
- case WindowExpression(function, spec) =>
- function.children.forall(eligibleToLiftUp) && eligibleToLiftUp(spec)
- case e => e.children.forall(eligibleToLiftUp)
- }
+ val newPlan = if (referencedAliases.isEmpty) {
+ p
+ } else {
+ val outerProjectList = collection.mutable.Seq(newProjectList: _*)
+ val innerProjectList =
+ collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
+ referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
+ outerProjectList.update(idx, alias.toAttribute)
+ innerProjectList += alias
}
- if (!aggregateExpressions.forall(eligibleToLiftUp)) {
- return agg
+ p.copy(
+ projectList = outerProjectList.toSeq,
+ child = Project(innerProjectList.toSeq, child)
+ )
+ }
+ newPlan.mapChildren(apply0)
+
+ case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
Review Comment:
Completely copied the original code except adding the final line to recursively applying on the new Project
--
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] anchovYu commented on a diff in pull request #42276: [SPARK-44714] Ease restriction of LCA resolution regarding queries with having
Posted by "anchovYu (via GitHub)" <gi...@apache.org>.
anchovYu commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1286515285
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,95 +131,97 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written bottom-up recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 and SPARK-44714 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case pOriginal: Project if ruleApplicableOnOperator(pOriginal, pOriginal.projectList)
+ && pOriginal.projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ val p @ Project(projectList, child) = pOriginal.mapChildren(apply0)
+ var aliasMap = AttributeMap.empty[AliasEntry]
+ val referencedAliases = collection.mutable.Set.empty[AliasEntry]
+ def unwrapLCAReference(e: NamedExpression): NamedExpression = {
+ e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
+ val aliasEntry = aliasMap.get(lcaRef.a).get
+ // If there is no chaining of lateral column alias reference, push down the alias
+ // and unwrap the LateralColumnAliasReference to the NamedExpression inside
+ // If there is chaining, don't resolve and save to future rounds
+ if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
+ referencedAliases += aliasEntry
+ lcaRef.ne
+ } else {
+ lcaRef
+ }
+ case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
+ // It shouldn't happen, but restore to unresolved attribute to be safe.
+ UnresolvedAttribute(lcaRef.nameParts)
+ }.asInstanceOf[NamedExpression]
+ }
+ val newProjectList = projectList.zipWithIndex.map {
+ case (a: Alias, idx) =>
+ val lcaResolved = unwrapLCAReference(a)
+ // Insert the original alias instead of rewritten one to detect chained LCA
+ aliasMap += (a.toAttribute -> AliasEntry(a, idx))
+ lcaResolved
+ case (e, _) =>
+ unwrapLCAReference(e)
+ }
- // Check if current Aggregate is eligible to lift up with Project: the aggregate
- // expression only contains: 1) aggregate functions, 2) grouping expressions, 3) leaf
- // expressions excluding attributes not in grouping expressions
- // This check is to prevent unnecessary transformation on invalid plan, to guarantee it
- // throws the same exception. For example, cases like non-aggregate expressions not
- // in group by, once transformed, will throw a different exception: missing input.
- def eligibleToLiftUp(exp: Expression): Boolean = {
- exp match {
- case _: AggregateExpression => true
- case e if groupingExpressions.exists(_.semanticEquals(e)) => true
- case a: Attribute => false
- case s: ScalarSubquery if s.children.nonEmpty
- && !groupingExpressions.exists(_.semanticEquals(s)) => false
- // Manually skip detection on function itself because it can be an aggregate function.
- // This is to avoid expressions like sum(salary) over () eligible to lift up.
- case WindowExpression(function, spec) =>
- function.children.forall(eligibleToLiftUp) && eligibleToLiftUp(spec)
- case e => e.children.forall(eligibleToLiftUp)
- }
- }
- if (!aggregateExpressions.forall(eligibleToLiftUp)) {
- return agg
+ if (referencedAliases.isEmpty) {
+ p
+ } else {
+ val outerProjectList = collection.mutable.Seq(newProjectList: _*)
+ val innerProjectList =
+ collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
+ referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
+ outerProjectList.update(idx, alias.toAttribute)
+ innerProjectList += alias
}
+ p.copy(
+ projectList = outerProjectList.toSeq,
+ child = Project(innerProjectList.toSeq, child)
+ )
+ }
+
+ case aggOriginal: Aggregate
+ if ruleApplicableOnOperator(aggOriginal, aggOriginal.aggregateExpressions)
+ && aggOriginal.aggregateExpressions.exists(
+ _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ val agg @ Aggregate(groupingExpressions, aggregateExpressions, _) =
+ aggOriginal.mapChildren(apply0)
+ // Check if current Aggregate is eligible to lift up with Project: the aggregate
+ // expression only contains: 1) aggregate functions, 2) grouping expressions, 3) leaf
+ // expressions excluding attributes not in grouping expressions
+ // This check is to prevent unnecessary transformation on invalid plan, to guarantee it
+ // throws the same exception. For example, cases like non-aggregate expressions not
+ // in group by, once transformed, will throw a different exception: missing input.
+ def eligibleToLiftUp(exp: Expression): Boolean = {
+ exp match {
+ case _: AggregateExpression => true
+ case e if groupingExpressions.exists(_.semanticEquals(e)) => true
+ case a: Attribute => false
+ case s: ScalarSubquery if s.children.nonEmpty
+ && !groupingExpressions.exists(_.semanticEquals(s)) => false
+ // Manually skip detection on function itself because it can be an aggregate function.
+ // This is to avoid expressions like sum(salary) over () eligible to lift up.
+ case WindowExpression(function, spec) =>
+ function.children.forall(eligibleToLiftUp) && eligibleToLiftUp(spec)
+ case e => e.children.forall(eligibleToLiftUp)
+ }
+ }
+ if (!aggregateExpressions.forall(eligibleToLiftUp)) {
+ agg
Review Comment:
Previously this line was `return agg`. That could be risky to `return` in a closure.. But for `apply0` with or without return should be the same.
--
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] anchovYu commented on a diff in pull request #42276: [WIP] Throw a more tailored error message for having - window - LCA unresolved case
Posted by "anchovYu (via GitHub)" <gi...@apache.org>.
anchovYu commented on code in PR #42276:
URL: https://github.com/apache/spark/pull/42276#discussion_r1285473731
##########
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/analysis/ResolveLateralColumnAliasReference.scala:
##########
@@ -131,138 +131,166 @@ object ResolveLateralColumnAliasReference extends Rule[LogicalPlan] {
(pList.exists(hasWindowExpression) && p.expressions.forall(_.resolved) && p.childrenResolved)
}
- override def apply(plan: LogicalPlan): LogicalPlan = {
- if (!conf.getConf(SQLConf.LATERAL_COLUMN_ALIAS_IMPLICIT_ENABLED)) {
- plan
- } else if (plan.containsAnyPattern(TEMP_RESOLVED_COLUMN, UNRESOLVED_HAVING)) {
- // It should not change the plan if `TempResolvedColumn` or `UnresolvedHaving` is present in
- // the query plan. These plans need certain plan shape to get recognized and resolved by other
- // rules, such as Filter/Sort + Aggregate to be matched by ResolveAggregateFunctions.
- // LCA resolution can break the plan shape, like adding Project above Aggregate.
- plan
- } else {
- // phase 2: unwrap
- plan.resolveOperatorsUpWithPruning(
- _.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE), ruleId) {
- case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
- && projectList.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
- var aliasMap = AttributeMap.empty[AliasEntry]
- val referencedAliases = collection.mutable.Set.empty[AliasEntry]
- def unwrapLCAReference(e: NamedExpression): NamedExpression = {
- e.transformWithPruning(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- case lcaRef: LateralColumnAliasReference if aliasMap.contains(lcaRef.a) =>
- val aliasEntry = aliasMap.get(lcaRef.a).get
- // If there is no chaining of lateral column alias reference, push down the alias
- // and unwrap the LateralColumnAliasReference to the NamedExpression inside
- // If there is chaining, don't resolve and save to future rounds
- if (!aliasEntry.alias.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) {
- referencedAliases += aliasEntry
- lcaRef.ne
- } else {
- lcaRef
- }
- case lcaRef: LateralColumnAliasReference if !aliasMap.contains(lcaRef.a) =>
- // It shouldn't happen, but restore to unresolved attribute to be safe.
- UnresolvedAttribute(lcaRef.nameParts)
- }.asInstanceOf[NamedExpression]
- }
- val newProjectList = projectList.zipWithIndex.map {
- case (a: Alias, idx) =>
- val lcaResolved = unwrapLCAReference(a)
- // Insert the original alias instead of rewritten one to detect chained LCA
- aliasMap += (a.toAttribute -> AliasEntry(a, idx))
- lcaResolved
- case (e, _) =>
- unwrapLCAReference(e)
- }
+ /** Internal application method. A hand-written top down recursive traverse. */
+ private def apply0(plan: LogicalPlan): LogicalPlan = {
+ plan match {
+ case p: LogicalPlan if !p.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE) =>
+ p
- if (referencedAliases.isEmpty) {
- p
- } else {
- val outerProjectList = collection.mutable.Seq(newProjectList: _*)
- val innerProjectList =
- collection.mutable.ArrayBuffer(child.output.map(_.asInstanceOf[NamedExpression]): _*)
- referencedAliases.foreach { case AliasEntry(alias: Alias, idx) =>
- outerProjectList.update(idx, alias.toAttribute)
- innerProjectList += alias
- }
- p.copy(
- projectList = outerProjectList.toSeq,
- child = Project(innerProjectList.toSeq, child)
- )
- }
+ // It should not change the Aggregate (and thus the plan shape) if its parent is an
+ // UnresolvedHaving, to avoid breaking the shape pattern `UnresolvedHaving - Aggregate`
+ // matched by ResolveAggregateFunctions. See SPARK-42936 for more details.
+ case u @ UnresolvedHaving(_, agg: Aggregate)
+ if agg.aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ u.copy(child = agg.mapChildren(apply0))
- case agg @ Aggregate(groupingExpressions, aggregateExpressions, _)
- if ruleApplicableOnOperator(agg, aggregateExpressions)
- && aggregateExpressions.exists(_.containsPattern(LATERAL_COLUMN_ALIAS_REFERENCE)) =>
+ case p @ Project(projectList, child) if ruleApplicableOnOperator(p, projectList)
Review Comment:
Completely copied the original code except adding the final line to recursively iterating children
--
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