Skip to content

[FLINK-38397][table] Fix stale sort trait on BatchPhysicalLocalHashAggregate causing ArrayIndexOutOfBoundsException #28004

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
Au-Miner wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[FLINK-38397][table] Fix stale sort trait on BatchPhysicalLocalHashAggregate causing ArrayIndexOutOfBoundsException #28004

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 4 additions & 2 deletions ...a/org/apache/flink/table/planner/plan/rules/physical/batch/BatchPhysicalHashAggRule.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,7 +30,7 @@ import org.apache.flink.table.planner.utils.TableConfigUtils.isOperatorDisabled

import org.apache.calcite.plan.{RelOptRule, RelOptRuleCall}
import org.apache.calcite.plan.RelOptRule.{any, operand}
import org.apache.calcite.rel.RelNode
import org.apache.calcite.rel.{RelCollations, RelNode}

import scala.collection.JavaConversions._

Expand Down Expand Up @@ -101,7 +101,9 @@ class BatchPhysicalHashAggRule
// create BatchPhysicalLocalHashAggregate
val localRequiredTraitSet = input.getTraitSet.replace(FlinkConventions.BATCH_PHYSICAL)
val newInput = RelOptRule.convert(input, localRequiredTraitSet)
val providedTraitSet = localRequiredTraitSet
// A hash aggregate does not preserve input ordering; reset the collation to EMPTY
// so the local agg does not advertise a stale collation from its input.
val providedTraitSet = localRequiredTraitSet.replace(RelCollations.EMPTY)
val localHashAgg = createLocalAgg(
agg.getCluster,
providedTraitSet,
Expand Down
13 changes: 13 additions & 0 deletions ...nner/src/test/resources/org/apache/flink/table/planner/plan/batch/table/AggregateTest.xml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -71,6 +71,19 @@ HashAggregate(isMerge=[true], select=[Final_AVG(sum$0, count$1) AS EXPR$0, Final
+- LocalHashAggregate(select=[Partial_AVG(a) AS (sum$0, count$1), Partial_SUM(b) AS sum$2, Partial_COUNT(c) AS count$3, Partial_SUM($f3) AS sum$4])
+- Calc(select=[CAST(1 AS INTEGER) AS a, b, c, c._1 AS $f3], where=[(a = 1)])
+- TableSourceScan(table=[[default_catalog, default_database, MyTable]], fields=[a, b, c])
]]>
</Resource>
</TestCase>
<TestCase name="testGroupAggregateAfterOrderBy">
<Resource name="optimized exec plan">
<![CDATA[
Calc(select=[EXPR$0 AS cnt])
+- HashAggregate(isMerge=[true], groupBy=[d], select=[d, Final_COUNT(count$0) AS EXPR$0])
+- Exchange(distribution=[hash[d]])
+- LocalHashAggregate(groupBy=[d], select=[d, Partial_COUNT(e) AS count$0])
+- Sort(orderBy=[c ASC])
+- Exchange(distribution=[single])
+- TableSourceScan(table=[[default_catalog, default_database, MyTable]], fields=[a, b, c, d, e])
]]>
</Resource>
</TestCase>
Expand Down
16 changes: 16 additions & 0 deletions ...lanner/src/test/scala/org/apache/flink/table/planner/plan/batch/table/AggregateTest.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,4 +73,20 @@ class AggregateTest extends TableTestBase {

util.verifyExecPlan(resultTable)
}

@Test
def testGroupAggregateAfterOrderBy(): Unit = {
// order_by before a group_by that prunes the sort-key column must not
// leave a stale collation on the local hash aggregate.
val util = batchTestUtil()
val sourceTable = util
.addTableSource[(Int, Long, Int, String, Double)]("MyTable", 'a, 'b, 'c, 'd, 'e)

val resultTable = sourceTable
.orderBy('c)
.groupBy('d)
.select('e.count.as('cnt))

util.verifyExecPlan(resultTable)
}
}