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

[fix](nereids)exists subquery should handle top level scarlar agg correctly (#29135)

Browse Source
This commit is contained in:
starocean999
2023-12-29 09:45:20 +08:00
committed by GitHub
parent 61677d1d4b
commit d2dc12bed5
3 changed files with 193 additions and 22 deletions
Download Patch File Download Diff File Expand all files Collapse all files

51
fe/fe-core/src/main/java/org/apache/doris/nereids/rules/analysis/SubqueryToApply.java
View File

@ -325,7 +325,10 @@ public class SubqueryToApply implements AnalysisRuleFactory {
LogicalPlan tmpPlan = childPlan;
for (int i = 0; i < subqueryExprs.size(); ++i) {
SubqueryExpr subqueryExpr = subqueryExprs.get(i);
if (nonMarkJoinExistsWithAgg(subqueryExpr, subqueryToMarkJoinSlot)) {
if (subqueryExpr instanceof Exists && hasTopLevelScalarAgg(subqueryExpr.getQueryPlan())) {
// because top level scalar agg always returns a value or null(for empty input)
// so Exists and Not Exists conjunct are always evaluated to True and false literals respectively
// we don't create apply node for it
continue;
}
@ -338,18 +341,11 @@ public class SubqueryToApply implements AnalysisRuleFactory {
return tmpPlan;
}
private boolean nonMarkJoinExistsWithAgg(SubqueryExpr exists,
Map<SubqueryExpr, Optional<MarkJoinSlotReference>> subqueryToMarkJoinSlot) {
return exists instanceof Exists
&& !subqueryToMarkJoinSlot.get(exists).isPresent()
&& hasTopLevelAggWithoutGroupBy(exists.getQueryPlan());
}
private static boolean hasTopLevelAggWithoutGroupBy(Plan plan) {
private static boolean hasTopLevelScalarAgg(Plan plan) {
if (plan instanceof LogicalAggregate) {
return ((LogicalAggregate) plan).getGroupByExpressions().isEmpty();
} else if (plan instanceof LogicalProject || plan instanceof LogicalSort) {
return hasTopLevelAggWithoutGroupBy(plan.child(0));
return hasTopLevelScalarAgg(plan.child(0));
}
return false;
}
@ -427,19 +423,30 @@ public class SubqueryToApply implements AnalysisRuleFactory {
// The result set when NULL is specified in the subquery and still evaluates to TRUE by using EXISTS
// When the number of rows returned is empty, agg will return null, so if there is more agg,
// it will always consider the returned result to be true
boolean needCreateMarkJoinSlot = isMarkJoin || shouldOutputMarkJoinSlot;
MarkJoinSlotReference markJoinSlotReference = null;
if (hasTopLevelAggWithoutGroupBy(exists.getQueryPlan()) && needCreateMarkJoinSlot) {
markJoinSlotReference =
new MarkJoinSlotReference(statementContext.generateColumnName(), true);
} else if (needCreateMarkJoinSlot) {
markJoinSlotReference =
new MarkJoinSlotReference(statementContext.generateColumnName());
if (hasTopLevelScalarAgg(exists.getQueryPlan())) {
/*
top level scalar agg and always return a value or null for empty input
so Exists and Not Exists conjunct are always evaluated to True and False literals respectively
SELECT *
FROM t1
WHERE EXISTS (
SELECT SUM(a)
FROM t2
WHERE t1.a = t2.b and t1.a = 1;
);
*/
return exists.isNot() ? BooleanLiteral.FALSE : BooleanLiteral.TRUE;
} else {
boolean needCreateMarkJoinSlot = isMarkJoin || shouldOutputMarkJoinSlot;
if (needCreateMarkJoinSlot) {
MarkJoinSlotReference markJoinSlotReference =
new MarkJoinSlotReference(statementContext.generateColumnName());
context.setSubqueryToMarkJoinSlot(exists, Optional.of(markJoinSlotReference));
return markJoinSlotReference;
} else {
return BooleanLiteral.TRUE;
}
}
if (needCreateMarkJoinSlot) {
context.setSubqueryToMarkJoinSlot(exists, Optional.of(markJoinSlotReference));
}
return needCreateMarkJoinSlot ? markJoinSlotReference : BooleanLiteral.TRUE;
}
@Override