fix: erase extract() bound symbol in UsedSymbolsCollector

[DavIvek]

Description

Closes #2023

A WHERE clause that uses extract(v IN list | expr) (and exercises the bound iterator inside expr) crashes the planner with the internal error:

Expected to generate all filters! Please contact Memgraph support as this scenario should not happen and is very likely a bug in the query engine!

Repro

CREATE (:Actor {domain:'test.com', infringes:[{brand:'Customer A'}]});

MATCH (i:Actor {domain:'test.com'})
WHERE 'Customer A' IN extract(v IN i.infringes | v.brand)
RETURN i.domain;
-- ↑ throws "Expected to generate all filters!"

The user-reported workaround (lifting extract(...) into a WITH and filtering on the alias) still works.

Root cause

UsedSymbolsCollector in src/query/plan/preprocess.hpp defines PostVisit overrides that erase the lambda-bound symbol for All, Any, Single, None, Reduce, and ListComprehension. The override for Extract was missing, so the iterator v of extract(v IN list | expr) ended up in the filter's used_symbols. Because no logical operator ever binds v, the filter is never matched against any expansion and PlanMatching ends with a non-empty filters set, tripping the guard in rule_based_planner.hpp:813.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[Copilot]

Pull request overview

Fixes a planner crash triggered by extract(v IN list | expr) used inside a WHERE clause by ensuring the lambda-bound iterator symbol (v) is treated as bound (not “free”) during symbol collection, so filter matching does not retain an unbindable symbol and trip an internal guard.

Changes:

• Add UsedSymbolsCollector::PostVisit(Extract&) to erase the iterator symbol bound by extract(...).
• Add new Gherkin/behave coverage reproducing the crash and validating correct behavior for matching and non-matching membership predicates.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
src/query/plan/preprocess.hpp	Erases extract’s bound iterator symbol from UsedSymbolsCollector results, aligning with existing handling for other lambda-like expressions.
tests/gql_behave/tests/memgraph_V1/features/issue_2023_extract_in_where.feature	Adds regression scenarios covering IN extract(...) in WHERE for both matching and non-matching cases.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[DavIvek]

Tracking

• [Link to Epic/Issue]

Standard development

• Update unit/E2E tests
• Compare the benchmarking results between the master branch this branch

CI Testing Labels

• Select the appropriate CI test labels (CI -build=build-name -test=test-suite)

Documentation checklist

• Add the documentation label
• Add the bug / feature label
• Add the milestone for which this feature is intended
  • If not known, set for a later milestone
• Write a release note, including added/changed clauses
  • Fixes a query planning crash when using extract() with a bound iterator inside a WHERE clause. Queries using this pattern now execute correctly without requiring users to rewrite them using an intermediate WITH. #4081