From the #2035 scenario review (private plan, reviews/pygraphistry-2035-hardening/scenarios.md). Two shapes decline to the full path today with parity, and are cheap to admit:
MATCH (m {id})-[:T]->(p) WHERE m.score > 0 RETURN p.age / WHERE p.age > 0 ... — a same-path WHERE whose predicates each touch one alias. The row pipeline already peels single-alias predicates into alias_prefilters; the seeded fast path could apply the same evaluator on its seed/destination rows before projecting (≈½ day incl. pins per predicate kind; IN lists are the same row).MATCH (p {id}) RETURN p LIMIT 1 — whole-row RETURN with a suffix; the props form already serves LIMIT/ORDER BY/DISTINCT through chain_impl(out, suffix_ops).
Not needed by the SNB shapes; backlog after the 0.60 stacks land.
From the #2035 scenario review (private plan,
reviews/pygraphistry-2035-hardening/scenarios.md). Two shapes decline to the full path today with parity, and are cheap to admit:MATCH (m {id})-[:T]->(p) WHERE m.score > 0 RETURN p.age/WHERE p.age > 0 ...— a same-path WHERE whose predicates each touch one alias. The row pipeline already peels single-alias predicates intoalias_prefilters; the seeded fast path could apply the same evaluator on its seed/destination rows before projecting (≈½ day incl. pins per predicate kind;INlists are the same row).MATCH (p {id}) RETURN p LIMIT 1— whole-row RETURN with a suffix; the props form already servesLIMIT/ORDER BY/DISTINCTthroughchain_impl(out, suffix_ops).Not needed by the SNB shapes; backlog after the 0.60 stacks land.