database/doris
2
0
Fork 0
Code Issues Pull Requests Actions 3 Packages Projects Releases Wiki Activity

[enhancement](nereids)eliminate repeat node if there is only 1 grouping set and no grouping scalar function (#35872)

Browse Source
This commit is contained in:
starocean999
2024-06-05 18:03:20 +08:00
committed by GitHub
parent af31e96c4e
commit bcde9c6524
3 changed files with 59 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
fe/fe-core/src/main/java/org/apache/doris/nereids/rules/analysis/NormalizeRepeat.java
View File

@ -82,6 +82,12 @@ public class NormalizeRepeat extends OneAnalysisRuleFactory {
public Rule build() {
return RuleType.NORMALIZE_REPEAT.build(
logicalRepeat(any()).when(LogicalRepeat::canBindVirtualSlot).then(repeat -> {
if (repeat.getGroupingSets().size() == 1
&& ExpressionUtils.collect(repeat.getOutputExpressions(),
GroupingScalarFunction.class::isInstance).isEmpty()) {
return new LogicalAggregate<>(repeat.getGroupByExpressions(),
repeat.getOutputExpressions(), repeat.child());
}
checkRepeatLegality(repeat);
repeat = removeDuplicateColumns(repeat);
// add virtual slot, LogicalAggregate and LogicalProject for normalize

39
fe/fe-core/src/test/java/org/apache/doris/nereids/rules/analysis/NormalizeRepeatTest.java
View File

@ -20,6 +20,7 @@ package org.apache.doris.nereids.rules.analysis;
import org.apache.doris.nereids.trees.expressions.Alias;
import org.apache.doris.nereids.trees.expressions.Slot;
import org.apache.doris.nereids.trees.expressions.functions.agg.Sum;
import org.apache.doris.nereids.trees.expressions.functions.scalar.GroupingId;
import org.apache.doris.nereids.trees.plans.Plan;
import org.apache.doris.nereids.trees.plans.logical.LogicalOlapScan;
import org.apache.doris.nereids.trees.plans.logical.LogicalRepeat;
@ -41,7 +42,7 @@ public class NormalizeRepeatTest implements MemoPatternMatchSupported {
Slot name = scan1.getOutput().get(1);
Alias alias = new Alias(new Sum(name), "sum(name)");
Plan plan = new LogicalRepeat<>(
ImmutableList.of(ImmutableList.of(id)),
ImmutableList.of(ImmutableList.of(id), ImmutableList.of(name)),
ImmutableList.of(idNotNull, alias),
scan1
);
@ -51,4 +52,40 @@ public class NormalizeRepeatTest implements MemoPatternMatchSupported {
logicalRepeat().when(repeat -> repeat.getOutputExpressions().get(0).nullable())
);
}
@Test
public void testEliminateRepeat() {
Slot id = scan1.getOutput().get(0);
Slot idNotNull = id.withNullable(true);
Slot name = scan1.getOutput().get(1);
Alias alias = new Alias(new Sum(name), "sum(name)");
Plan plan = new LogicalRepeat<>(
ImmutableList.of(ImmutableList.of(id)),
ImmutableList.of(idNotNull, alias),
scan1
);
PlanChecker.from(MemoTestUtils.createCascadesContext(plan))
.applyTopDown(new NormalizeRepeat())
.matchesFromRoot(
logicalAggregate(logicalOlapScan())
);
}
@Test
public void testNoEliminateRepeat() {
Slot id = scan1.getOutput().get(0);
Slot idNotNull = id.withNullable(true);
Slot name = scan1.getOutput().get(1);
Alias alias = new Alias(new GroupingId(name), "grouping_id(name)");
Plan plan = new LogicalRepeat<>(
ImmutableList.of(ImmutableList.of(id)),
ImmutableList.of(idNotNull, alias),
scan1
);
PlanChecker.from(MemoTestUtils.createCascadesContext(plan))
.applyTopDown(new NormalizeRepeat())
.matchesFromRoot(
logicalAggregate(logicalRepeat(logicalOlapScan()))
);
}
}

15
regression-test/suites/nereids_rules_p0/grouping_sets/grouping_normalize_test.groovy
View File

@ -39,4 +39,19 @@ suite("grouping_normalize_test"){
SELECT ROUND( SUM(pk + 1) - 3) col_alias1, MAX( DISTINCT col_int_undef_signed - 5) AS col_alias2, pk + 1 AS col_alias3
FROM grouping_normalize_test GROUP BY GROUPING SETS ((col_int_undef_signed,col_int_undef_signed2,pk),()) order by 1,2,3;
"""
explain {
sql("SELECT col_int_undef_signed, col_int_undef_signed2, SUM(pk) FROM grouping_normalize_test GROUP BY GROUPING SETS ((col_int_undef_signed, col_int_undef_signed2));")
notContains("VREPEAT_NODE")
}
explain {
sql("SELECT col_int_undef_signed, col_int_undef_signed2, SUM(pk), grouping_id(col_int_undef_signed2) FROM grouping_normalize_test GROUP BY GROUPING SETS ((col_int_undef_signed, col_int_undef_signed2),());")
contains("VREPEAT_NODE")
}
explain {
sql("SELECT col_int_undef_signed, col_int_undef_signed2, SUM(pk) FROM grouping_normalize_test GROUP BY GROUPING SETS ((col_int_undef_signed, col_int_undef_signed2));")
notContains("VREPEAT_NODE")
}
}