Repository navigation
branch-4.2: [fix](rbo) Give the constants materialized by constant propagation a fresh exprId #69111 - #69406
Merged
Conversation
…fresh exprId (#69111) ### What problem does this PR solve? Problem Summary: When ConstantPropagation replaces an aggregate output by a constant, it materializes the constant in a project above the aggregate, and the project item reused the exprId of the replaced output: LogicalProject ( projects=[..., 1 AS `id`#1, ...] ) +-- LogicalAggregate ( groupByExpr=[id#1], outputExpr=[id#1] ) So one exprId had two definitions, the aggregate output slot and the constant, whose nullable and data type may differ. Rules which compare exprIds only then take them as the same column. For RQG case 11000258898 EliminateUnnecessaryProject removed such a project after the aggregate had been rewritten into `Limit 1`, the plan above fell back to the nullable output slot of the set operation, and the last AdjustNullable pass reported AdjustNullable convert slot 2004-04-28#55 from not-nullable to nullable Fix: Give the constant a fresh exprId only when it really shadows a column, i.e. when the exprId is still produced by the plan below; otherwise keep it: * the constant gets a fresh exprId, and the references above are refreshed with ExprIdRewriter, when the aggregate still outputs the replaced column (its grouping key is kept because all grouping keys are constant) or when a project item re-defines a column which its child still outputs. That is the shadow the old code created, e.g. the RQG case 11000258898 of DORIS-28416. * the constant keeps the exprId when the aggregate removed the constant grouping key from its outputs: the project item is then the only definition of the column, nothing below references the exprId, and consumers which rely on the identity of the column, e.g. the materialized view matching, keep working. The plan of such a query is byte identical to the plan before this change, including the exprIds (verified with the query of `agg_optimize_when_uniform/mv2_0`, for the query side and for the MV side which is generated with constant propagation disabled). The sink also has to follow the fresh exprIds, otherwise it would reference a slot which its child does not output any more; only exprIds are remapped there, the name and the data type declared by the sink are kept (that is why the rule does not replace the slots of the sink by literals). A CTE is rewritten in separate subtree contexts (producer body and consumers) and the consumers, which map their slots to the producer outputs by exprId, are rewritten before the producer. The renamed exprIds are published per CTE (`StatementContext#producerExprIdRemap`), and `RewriteCteChildren` refreshes both maps of every consumer with the last exprId of the chain (`ExprIdRewriter#lastExprId`). Reproduction: ConstantPropagationFreshExprIdTest runs this rule alone, so the project above the aggregate is still visible in the plan, and asserts that it does not reuse the exprId of the aggregate output and that the plan above references the fresh exprIds. ConstantPropagationTest used to pin the reused exprIds in its constant project expectations; it now asserts fresh exprIds. Both fail before this change. ### Release note None ### Check List (For Author) - Test: Unit Test - ConstantPropagationFreshExprIdTest (new): fails before with "the constant project reuses the exprId of the aggregate output". - ConstantPropagationTest: expectations updated, fails before with "the constant project item must not reuse the exprId of the replaced output slot". - 31 related FE unit test classes (constant propagation, constant cte, set operation, column pruning, fd, uniform, sql plan suite, runtime filter, ...) rerun: 385 tests, 0 failures. - Behavior changed: Yes. A constant materialized above an aggregate has its own exprId and the plan above references it, so one exprId no longer describes two different columns. - Does this need documentation: No
Contributor
|
run buildall |
Contributor
|
Thank you for your contribution to Apache Doris. Please clearly describe your PR:
|
Contributor
|
run buildall |
The auto pick brought the master test files over verbatim and two of them do not compile on branch-4.2: - ExprIdRewriterTest#testGuardedBranchIsRewritten builds its assignment with MergeUtils.selectBranch, which returns a ShortCircuitIf. branch-4.2's MergeUtils only has buildMergeJoin and there is no ShortCircuitIf class (MERGE INTO does not wrap its assignments that way here), so the subtree that fix made reachable does not exist on this branch: the test is dropped, the structural exprId replacement it exercises is ported unchanged. - The two plan-rewrite tests of ExprIdRewriterTest use the 2-arg JobContext constructor; branch-4.2 has the 3-arg one (ScheduleContext, PhysicalProperties, double). Also drop the two RQG reproductions of ConstantPropagationFreshExprIdTest: both queries already trip "AdjustNullable convert slot 2004-04-28#55 from not-nullable to nullable" (and "cKToNmsxok#45") on the pristine branch-4.2, i.e. before this change, and the failure is unchanged by it. ConstantPropagation does not modify the plans of those queries on this branch at all (verified by dumping the plan with the rule applied alone), so the failure comes from another rule here, while on master the rule does rewrite them. The tables those queries need are still created by the test. Verified on branch-4.2: 105 tests of the 15 related FE unit test classes pass, and the nereids_rules_p0/constant_propagation and nereids_rules_p0/mv/agg_optimize_when_uniform regression suites pass.
Contributor
|
run buildall |
Contributor
FE UT Coverage ReportIncrement line coverage |
Contributor
FE Regression Coverage ReportIncrement line coverage |
Contributor
|
skip buildall |
yiguolei
approved these changes
Oct 23, 2026
Contributor
Author
|
PR approved by anyone and no changes requested. |
github-actions
Bot
added
the
approved
Contributor
Author
|
PR approved by at least one committer and no changes requested. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Cherry-picked from #69111