This is an automated email from the ASF dual-hosted git repository.

yiguolei pushed a commit to branch branch-4.1
in repository https://gitbox.apache.org/repos/asf/doris.git


The following commit(s) were added to refs/heads/branch-4.1 by this push:
     new 75801240671 branch-4.1 [fix](nereids)Fix CTE consumer stats derivation 
when producer stats are not precomputed #61074 (#61473)
75801240671 is described below

commit 758012406716e294f7a1bdb64ecfa22d25bef93c
Author: minghong <[email protected]>
AuthorDate: Thu Mar 19 14:07:48 2026 +0800

    branch-4.1 [fix](nereids)Fix CTE consumer stats derivation when producer 
stats are not precomputed #61074 (#61473)
    
    ### What problem does this PR solve?
    pick #61074
    Issue Number: close #xxx
    
    Related PR: #xxx
    
    Problem Summary:
    
    ### Release note
    
    None
    
    ### Check List (For Author)
    
    - Test <!-- At least one of them must be included. -->
        - [ ] Regression test
        - [ ] Unit Test
        - [ ] Manual test (add detailed scripts or steps below)
        - [ ] No need to test or manual test. Explain why:
    - [ ] This is a refactor/code format and no logic has been changed.
            - [ ] Previous test can cover this change.
            - [ ] No code files have been changed.
            - [ ] Other reason <!-- Add your reason?  -->
    
    - Behavior changed:
        - [ ] No.
        - [ ] Yes. <!-- Explain the behavior change -->
    
    - Does this need documentation?
        - [ ] No.
    - [ ] Yes. <!-- Add document PR link here. eg:
    https://github.com/apache/doris-website/pull/1214 -->
    
    ### Check List (For Reviewer who merge this PR)
    
    - [ ] Confirm the release note
    - [ ] Confirm test cases
    - [ ] Confirm document
    - [ ] Add branch pick label <!-- Add branch pick label that this PR
    should merge into -->
---
 .../org/apache/doris/nereids/StatementContext.java | 23 +++++++++++-----------
 .../doris/nereids/rules/analysis/AnalyzeCTE.java   |  4 +++-
 .../rewrite/DecomposeRepeatWithPreAggregation.java |  1 +
 .../doris/nereids/rules/rewrite/OrExpansion.java   |  2 ++
 .../nereids/rules/rewrite/RewriteCteChildren.java  |  5 ++++-
 .../rules/rewrite/SplitMultiDistinctStrategy.java  |  4 +---
 .../doris/nereids/rules/rewrite/StatsDerive.java   | 15 +++++++++++---
 7 files changed, 35 insertions(+), 19 deletions(-)

diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/StatementContext.java 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/StatementContext.java
index b3db083dc2a..b86f0b1eebc 100644
--- a/fe/fe-core/src/main/java/org/apache/doris/nereids/StatementContext.java
+++ b/fe/fe-core/src/main/java/org/apache/doris/nereids/StatementContext.java
@@ -55,6 +55,7 @@ import org.apache.doris.nereids.trees.plans.Plan;
 import org.apache.doris.nereids.trees.plans.RelationId;
 import org.apache.doris.nereids.trees.plans.TableId;
 import org.apache.doris.nereids.trees.plans.logical.LogicalCTEConsumer;
+import org.apache.doris.nereids.trees.plans.logical.LogicalCTEProducer;
 import org.apache.doris.nereids.trees.plans.logical.LogicalPlan;
 import org.apache.doris.nereids.util.RelationUtil;
 import org.apache.doris.qe.ConnectContext;
@@ -158,7 +159,7 @@ public class StatementContext implements Closeable {
     private final Map<CTEId, Set<LogicalCTEConsumer>> cteIdToConsumers = new 
HashMap<>();
     private final Map<CTEId, Set<Slot>> cteIdToOutputIds = new HashMap<>();
 
-    private final Map<CTEId, Statistics> cteIdToProducerStats = new 
HashMap<>();
+    private final Map<CTEId, LogicalCTEProducer<? extends Plan>> 
cteIdToProducer = new HashMap<>();
 
     private final Map<RelationId, Set<Expression>> consumerIdToFilters = new 
HashMap<>();
     // Used to update consumer's stats
@@ -644,7 +645,7 @@ public class StatementContext implements Closeable {
         return new CteEnvironmentSnapshot(
                 copyMapOfSets(cteIdToConsumers),
                 copyMapOfSets(cteIdToOutputIds),
-                new HashMap<>(cteIdToProducerStats),
+                new HashMap<>(cteIdToProducer),
                 copyMapOfSets(consumerIdToFilters),
                 copyMapOfLists(cteIdToConsumerGroup),
                 new HashMap<>(rewrittenCteProducer),
@@ -659,8 +660,8 @@ public class StatementContext implements Closeable {
         cteIdToOutputIds.clear();
         cteIdToOutputIds.putAll(snapshot.cteIdToOutputIds);
 
-        cteIdToProducerStats.clear();
-        cteIdToProducerStats.putAll(snapshot.cteIdToProducerStats);
+        cteIdToProducer.clear();
+        cteIdToProducer.putAll(snapshot.cteIdToProducer);
 
         consumerIdToFilters.clear();
         consumerIdToFilters.putAll(snapshot.consumerIdToFilters);
@@ -695,7 +696,7 @@ public class StatementContext implements Closeable {
     public static class CteEnvironmentSnapshot {
         private final Map<CTEId, Set<LogicalCTEConsumer>> cteIdToConsumers;
         private final Map<CTEId, Set<Slot>> cteIdToOutputIds;
-        private final Map<CTEId, Statistics> cteIdToProducerStats;
+        private final Map<CTEId, LogicalCTEProducer<? extends Plan>> 
cteIdToProducer;
         private final Map<RelationId, Set<Expression>> consumerIdToFilters;
         private final Map<CTEId, List<Pair<Multimap<Slot, Slot>, Group>>> 
cteIdToConsumerGroup;
         private final Map<CTEId, LogicalPlan> rewrittenCteProducer;
@@ -707,14 +708,14 @@ public class StatementContext implements Closeable {
         public CteEnvironmentSnapshot(
                 Map<CTEId, Set<LogicalCTEConsumer>> cteIdToConsumers,
                 Map<CTEId, Set<Slot>> cteIdToOutputIds,
-                Map<CTEId, Statistics> cteIdToProducerStats,
+                Map<CTEId, LogicalCTEProducer<? extends Plan>> cteIdToProducer,
                 Map<RelationId, Set<Expression>> consumerIdToFilters,
                 Map<CTEId, List<Pair<Multimap<Slot, Slot>, Group>>> 
cteIdToConsumerGroup,
                 Map<CTEId, LogicalPlan> rewrittenCteProducer,
                 Map<CTEId, LogicalPlan> rewrittenCteConsumer) {
             this.cteIdToConsumers = cteIdToConsumers;
             this.cteIdToOutputIds = cteIdToOutputIds;
-            this.cteIdToProducerStats = cteIdToProducerStats;
+            this.cteIdToProducer = cteIdToProducer;
             this.consumerIdToFilters = consumerIdToFilters;
             this.cteIdToConsumerGroup = cteIdToConsumerGroup;
             this.rewrittenCteProducer = rewrittenCteProducer;
@@ -1117,12 +1118,12 @@ public class StatementContext implements Closeable {
         return prepareStage;
     }
 
-    public Statistics getProducerStatsByCteId(CTEId id) {
-        return cteIdToProducerStats.get(id);
+    public LogicalCTEProducer<? extends Plan> getCteProducerByCteId(CTEId id) {
+        return cteIdToProducer.get(id);
     }
 
-    public void setProducerStats(CTEId id, Statistics stats) {
-        cteIdToProducerStats.put(id, stats);
+    public void setCteProducer(CTEId id, LogicalCTEProducer<? extends Plan> 
producer) {
+        cteIdToProducer.put(id, producer);
     }
 
     public void setIsInsert(boolean isInsert) {
diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/analysis/AnalyzeCTE.java
 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/analysis/AnalyzeCTE.java
index 4287e8ca7de..c49b3f01c51 100644
--- 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/analysis/AnalyzeCTE.java
+++ 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/analysis/AnalyzeCTE.java
@@ -109,7 +109,9 @@ public class AnalyzeCTE extends OneAnalysisRuleFactory {
                     aliasQuery.withChildren(ImmutableList.of(analyzedCtePlan));
             outerCteCtx = new CTEContext(cteId, logicalSubQueryAlias, 
outerCteCtx);
             outerCteCtx.setAnalyzedPlan(logicalSubQueryAlias);
-            cteProducerPlans.add(new LogicalCTEProducer<>(cteId, 
logicalSubQueryAlias));
+            LogicalCTEProducer<Plan> cteProducer = new 
LogicalCTEProducer<>(cteId, logicalSubQueryAlias);
+            cascadesContext.getStatementContext().setCteProducer(cteId, 
cteProducer);
+            cteProducerPlans.add(cteProducer);
         }
         return Pair.of(outerCteCtx, cteProducerPlans);
     }
diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/DecomposeRepeatWithPreAggregation.java
 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/DecomposeRepeatWithPreAggregation.java
index db97df41fe2..be4066fe221 100644
--- 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/DecomposeRepeatWithPreAggregation.java
+++ 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/DecomposeRepeatWithPreAggregation.java
@@ -492,6 +492,7 @@ public class DecomposeRepeatWithPreAggregation extends 
DefaultPlanRewriter<Disti
         }
         LogicalCTEProducer<LogicalAggregate<Plan>> producer =
                 new LogicalCTEProducer<>(ctx.statementContext.getNextCTEId(), 
preAggClone);
+        ctx.statementContext.setCteProducer(producer.getCteId(), producer);
         ctx.cteProducerList.add(producer);
         producer.accept(new StatsDerive(false), new DeriveContext());
         return producer;
diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/OrExpansion.java
 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/OrExpansion.java
index 4cb07fd44fd..2048ea8f50e 100644
--- 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/OrExpansion.java
+++ 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/OrExpansion.java
@@ -141,10 +141,12 @@ public class OrExpansion extends 
DefaultPlanRewriter<OrExpandsionContext> implem
                 .deepCopy((LogicalPlan) join.left(), new DeepCopierContext());
         LogicalCTEProducer<? extends Plan> leftProducer = new 
LogicalCTEProducer<>(
                 ctx.statementContext.getNextCTEId(), leftClone);
+        ctx.statementContext.setCteProducer(leftProducer.getCteId(), 
leftProducer);
         LogicalPlan rightClone = LogicalPlanDeepCopier.INSTANCE
                 .deepCopy((LogicalPlan) join.right(), new DeepCopierContext());
         LogicalCTEProducer<? extends Plan> rightProducer = new 
LogicalCTEProducer<>(
                 ctx.statementContext.getNextCTEId(), rightClone);
+        ctx.statementContext.setCteProducer(rightProducer.getCteId(), 
rightProducer);
         Map<Slot, Slot> leftCloneToLeft = new HashMap<>();
         for (int i = 0; i < leftClone.getOutput().size(); i++) {
             leftCloneToLeft.put(leftClone.getOutput().get(i), 
(join.left()).getOutput().get(i));
diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/RewriteCteChildren.java
 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/RewriteCteChildren.java
index 31aeb0be1ba..0360c4aa0e7 100644
--- 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/RewriteCteChildren.java
+++ 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/RewriteCteChildren.java
@@ -147,7 +147,10 @@ public class RewriteCteChildren extends 
DefaultPlanRewriter<CascadesContext> imp
             cascadesContext.addPlanProcesses(rewrittenCtx.getPlanProcesses());
             
cascadesContext.getStatementContext().getRewrittenCteProducer().put(cteProducer.getCteId(),
 child);
         }
-        return cteProducer.withChildren(child);
+        LogicalCTEProducer<? extends Plan> rewrittenProducer = 
(LogicalCTEProducer<? extends Plan>) cteProducer
+                .withChildren(child);
+        
cascadesContext.getStatementContext().setCteProducer(rewrittenProducer.getCteId(),
 rewrittenProducer);
+        return rewrittenProducer;
     }
 
     private LogicalPlan pushPlanUnderAnchor(LogicalPlan plan) {
diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/SplitMultiDistinctStrategy.java
 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/SplitMultiDistinctStrategy.java
index 8fc150c7055..c9585d269aa 100644
--- 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/SplitMultiDistinctStrategy.java
+++ 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/SplitMultiDistinctStrategy.java
@@ -18,7 +18,6 @@
 package org.apache.doris.nereids.rules.rewrite;
 
 import 
org.apache.doris.nereids.rules.rewrite.DistinctAggStrategySelector.DistinctSelectorContext;
-import org.apache.doris.nereids.rules.rewrite.StatsDerive.DeriveContext;
 import org.apache.doris.nereids.trees.copier.DeepCopierContext;
 import org.apache.doris.nereids.trees.copier.LogicalPlanDeepCopier;
 import org.apache.doris.nereids.trees.expressions.Alias;
@@ -60,9 +59,8 @@ public class SplitMultiDistinctStrategy {
                 .deepCopy(agg, new DeepCopierContext());
         LogicalCTEProducer<Plan> producer = new 
LogicalCTEProducer<>(ctx.statementContext.getNextCTEId(),
                 cloneAgg.child());
+        ctx.statementContext.setCteProducer(producer.getCteId(), producer);
         ctx.cteProducerList.add(producer);
-        StatsDerive derive = new StatsDerive(false);
-        producer.accept(derive, new DeriveContext());
 
         Map<Slot, Slot> originToProducerSlot = new HashMap<>();
         for (int i = 0; i < agg.child().getOutput().size(); ++i) {
diff --git 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/StatsDerive.java
 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/StatsDerive.java
index e3e06c23e6b..7ad4e30bb22 100644
--- 
a/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/StatsDerive.java
+++ 
b/fe/fe-core/src/main/java/org/apache/doris/nereids/rules/rewrite/StatsDerive.java
@@ -351,20 +351,29 @@ public class StatsDerive extends PlanVisitor<Statistics, 
StatsDerive.DeriveConte
 
     @Override
     public Statistics visitLogicalCTEProducer(LogicalCTEProducer<? extends 
Plan> cteProducer, DeriveContext context) {
+        // Fallback registration: ensure cteId -> producer mapping is always 
available
+        // even if some upstream rewrite path misses explicit registration.
+        
ConnectContext.get().getStatementContext().setCteProducer(cteProducer.getCteId(),
 cteProducer);
         Statistics prodStats = cteProducer.child().accept(this, context);
         StatisticsBuilder builder = new StatisticsBuilder(prodStats);
         builder.setWidthInJoinCluster(1);
         Statistics stats = builder.build();
         cteProducer.setStatistics(stats);
-        
ConnectContext.get().getStatementContext().setProducerStats(cteProducer.getCteId(),
 stats);
         return stats;
     }
 
     @Override
     public Statistics visitLogicalCTEConsumer(LogicalCTEConsumer cteConsumer, 
DeriveContext context) {
         CTEId cteId = cteConsumer.getCteId();
-        Statistics prodStats = 
ConnectContext.get().getStatementContext().getProducerStatsByCteId(cteId);
-        Preconditions.checkArgument(prodStats != null, String.format("Stats 
for CTE: %s not found", cteId));
+        LogicalCTEProducer<? extends Plan> cteProducer = 
ConnectContext.get().getStatementContext()
+                .getCteProducerByCteId(cteId);
+        Preconditions.checkState(cteProducer != null,
+                String.format("CTE producer for CTE: %s not found", cteId));
+        Statistics prodStats = cteProducer.getStats();
+        if (prodStats == null || deepDerive) {
+            prodStats = cteProducer.accept(this, context);
+        }
+        Preconditions.checkState(prodStats != null, String.format("Stats for 
CTE: %s not found", cteId));
         Statistics consumerStats = new Statistics(prodStats.getRowCount(), 1, 
new HashMap<>());
         for (Slot slot : cteConsumer.getOutput()) {
             Slot prodSlot = cteConsumer.getProducerSlot(slot);


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to