Repository navigation
feat(gfql): preserve collected node identity through reentry - #2261
Merged
Merged
Conversation
lmeyerov
force-pushed
the
fix/gfql-2252-scalar-carry
branch
from
November 1, 2026 00:37
e131383 to
ee08a09
Compare
lmeyerov
force-pushed
the
work/2249-stage2-identity-polars
branch
from
November 1, 2026 02:54
0121abb to
398a684
Compare
lmeyerov
force-pushed
the
work/2249-stage2-identity-polars
branch
from
November 1, 2026 07:48
2d403ba to
7c5d405
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Stage 2 of #2249. Collected nodes keep their identity through
WITH collect(node) ... MATCH ... WHERE candidate IN nodesre-entry on pandas, Polars and cuDF. Grouped whole-entity keys (RETURN b, count(*),RETURN rel, count(*)) also stay identity-distinct instead of collapsing to null.Engine.null_column: Polars and cuDF are dtype-preserving, and pandas keeps the objectNonesentinel for parity.entity_props._false_maskbuilds its all-False mask fromisna(). The old(base == True) & FalseraisedArrowTypeErroron cuDF integer id columns.Rebased onto master
5819f46bafter #2258 landed. The duplicate #2258 cherry-pick is gone.Evidence at head
b9dc6b6bgfql-routes-offmodes with empty divergence ledgers. RTD is green.graphistry/test-rapids-official:26.02-gfql(cuDF 26.02) with an identity-asserted runner. The set is all 100 cuDF and Polars-GPU ids collected from the 11 touched test files withTEST_CUDF=1 TEST_POLARS_GPU=1. Skips count as failures. Result: 98 passed, 2 failed, 0 skipped.5819f46bin the same image:test_row_pipeline_order_by_stringified_list_subscript_expression_on_cudf_when_availabletest_row_pipeline_cudf_ast_map_literal_vector_values_when_available__array__conversion.test_append_property_segments_integer_alias_typed_false_mask[pandas,cudf]passes on both engines.Follow-ups
Engine.py.work/2249-polars-alias-width-pin: Polars named-node width pin, to open as its own PR.Part of #2249.
🤖 Generated with Claude Code