Untitled conversation
- User
[SQL] "expand" projection is always named "value" ignoring "AS"
ArcadeDB Server v26.5.1 (build b6d08fee549dd5930303b362223869901aab0fac/1778533954349/main)Running on Linux 6.17.0-29-generic - OpenJDK 64-Bit Server VM 25.0.3-eaUsing
expandin a projection, the result-set contains an property namedvalueinspite an alias of the formAS x:SELECT expand([1,2,3,4]) AS testresults in:
{"result":[{"value":1},{"value":2},{"value":3},{"value":4}]}Interface
No new interfaces are introduced.
- Assistant
I’ll trace the SQL projection/expand execution path, add a focused regression test, implement the alias propagation, and run the narrow test suite plus any nearby tests affected.
- Assistant
rgisn’t installed in this checkout, so I’m switching to targetedfind/grepsearches and keeping the scope to SQL execution and tests. - Assistant
The failure is localized: planning strips the explicit alias from the
expand()projection, then the expansion step hard-codes scalar rows to"value". I’m checking planner-copy behavior before choosing the smallest fix that preserves existing unaliased and record/map expansion semantics. - Assistant
The implementation now carries the explicit alias through planning only for scalar expanded values; unaliased scalar expansion still uses
"value", while maps/documents keep their existing flattened fields. I’ve added regression coverage for both aliased and unaliased cases and am running the focused engine test. - Assistant
The initial Maven invocation attempted to resolve its test provider from Central, so I stopped it and am rerunning Maven in offline mode to keep verification strictly local. Compilation had already completed successfully.
- Assistant
The focused Maven test runner is locally blocked because its Surefire provider JAR is absent from the cache. The project’s compile and test-compile phases pass; I’m now invoking the already-cached JUnit Platform directly against the compiled focused test, still offline.
- Assistant
The focused regression test passes: 8/8 tests in
ExpandParseTest. I also kept the new alias-aware constructor package-private so this does not expand the public API, and I’m doing one final compile plus the neighboringExpandStepTestsuite.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
- composer-2.5Compare side by side →
- composer-2.5Compare side by side →
- composer-2.5Compare side by side →
- composer-2.5Compare side by side →
- composer-2.5Compare side by side →
- gpt-5.6-solCompare side by side →
- gpt-5.6-solCompare side by side →
- gpt-5.6-solCompare side by side →
- gpt-5.6-solCompare side by side →
- fable-5Compare side by side →
- fable-5Compare side by side →