BE-597: Array-based type filtering in the SelectCompiler

[TimDiekmann]

🌟 What is the purpose of this PR?

Type filters (type.versionedUrl / type.baseUrl without an inheritance depth) compiled to one entity_is_of_type → ontology_temporal_metadata → ontology_ids join chain per predicate. On the trace that started this investigation (7 type exclusions), that meant 7 multiplying join groups, a planner misestimate, and a row-explosion the query had to DISTINCT away again.

This PR compiles those filters to array predicates (@> / &&) on the entity_edition_cache type arrays (from #8854) instead. Equality filters in the same All/Any group bundle into a single predicate over one parameter array, which the GIN index on the cache columns can serve.

🔗 Related links

• BE-597

🚫 Blocked by

• BE-599: Compute entity summary aggregations in SQL #8864

🔍 What does this change?

• BinaryOperator::ArrayContains (@>) in the expression AST; && (overlap) already existed.
• Path mapping (query/entity.rs): EntityTypeEdge { IsOfType, BaseUrl | VersionedUrl, inheritance_depth: None } resolves directly to the entity_edition_cache array columns. Explicit inheritance depths keep the join chain (the cache arrays are all-depth).
• The redundant EntityQueryPath::TypeBaseUrls / TypeVersionedUrls variants are deleted; the edge form is the single canonical path (they were documented as not queryable from the API, so no public surface changes).
• Compiler rewrite (query/compile.rs), type-driven via the column's declared ParameterType rather than hardcoded paths:
  • Equal/NotEqual/In(parameter, path) on text-array columns compile to containment predicates.
  • All/Any groups bundle same-column equality filters: All+equals → @>, Any+equals → &&, All+notEquals → NOT(&&), Any+notEquals → NOT(@>) — each an exact logical equivalent of the unbundled forms.
  • Bundles are keyed on the aliased column, so an entity's own type and a linked entity's type (same column, different join chain) stay separate predicates.
  • StartsWith/EndsWith/ContainsSegment on array-backed paths fail at compile time (UnsupportedTextArrayOperation) instead of producing a Postgres type error.
• Filter::for_entity_by_type_id builds a single versionedUrl equality instead of All[baseUrl, version] (the old shape was only correct through shared join aliases). All IsOfType/IsOfBaseType policy filters now go through the array path.
• The property-protection filters (In(param, type base URLs)) compile to base_urls @> ARRAY[$n] instead of $n = ANY(base_urls) — semantically equivalent, but GIN-indexable.

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

This PR:

• does not modify any publishable blocks or libraries, or modifications do not need publishing

📜 Does this require a change to the docs?

The changes in this PR:

• are internal and do not require a docs change

🕸️ Does this require a change to the Turbo Graph?

The changes in this PR:

• do not affect the execution graph

⚠️ Known issues

• Intentional behaviour change: notEqual on type paths now uses set semantics — "the entity has no such type" — instead of the old per-join-row semantics, under which a multi-typed entity matched notEqual(A) as long as it had any other type. The old behaviour made type exclusions effectively useless for multi-typed entities (an entity with types {notification, X} stayed visible despite a notEqual(notification) filter); the new behaviour is what the filters are meant to express. The same set semantics applies to separate baseUrl/version equality filters within one all block.
• Nested paths through link edges (leftEntity.type.versionedUrl etc.) intentionally keep the join chain for the link traversal; only the terminal type lookup uses the cache.

🐾 Next steps

• BE-598: keys-then-hydrate split (the remaining cost is the label-sort scan, not the type predicates).
• Integration test pinning the notEqual exclusion semantics on a multi-typed entity.
• Filter normalization pass (flattening nested All/Any, sibling dedup) to widen bundling across artificial nesting and improve planner estimates — separate ticket.

🛡 What tests cover this?

• New compiler snapshot tests in statement/select.rs: single equality (constructor + raw filter forms), Any/All bundling, bundled NOT(&&) exclusion, base-URL policy form, and the own-type vs. linked-type alias-separation regression test.
• Updated property-masking snapshot (= ANY → @>).
• email_filter_protection.rs integration tests cover the rewritten protection-filter path in CI.
• Manually validated against an 8.8M-entity dataset: counts and all summary maps byte-identical to the join-based compilation.

❓ How to test this?

1. Checkout the branch, start the graph with a seeded database.
2. Run a queryEntities request with type filters (equal/notEqual on ["type", "versionedUrl"], with and without an any/all group).
3. Confirm the compiled SQL (debug logs) contains entity_edition_cache array predicates instead of entity_is_of_type joins, and results match the base branch — except multi-typed entities under notEqual, which are now correctly excluded.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

3 Skipped Deployments

Project	Deployment	Actions	Updated (UTC)
hash	Ignored	Preview	Jun 12, 2026 6:54pm
hashdotdesign-tokens	Ignored	Preview	Jun 12, 2026 6:54pm
petrinaut	Skipped		Jun 12, 2026 6:54pm

[cursor]

PR Summary

Medium Risk
Changes core query compilation and filter semantics for type exclusions on multi-typed entities; behavior is intentional but can alter production query results versus the join-based compiler.

Overview
Entity type filters now hit entity_edition_cache array columns instead of one entity_is_of_type join chain per predicate. The Postgres SelectCompiler recognizes equality/In on materialized text[] type paths and emits @> / && (and negated forms), with All/Any bundling multiple type equalities into a single array predicate so GIN indexes can be used.

EntityQueryPath::TypeBaseUrls / TypeVersionedUrls are removed in favor of EntityTypeEdge { IsOfType, BaseUrl | VersionedUrl, inheritance_depth: None } for selections, summaries, and property protection. Filter::for_entity_by_type_id is a single versionedUrl equality instead of All[baseUrl, version].

notEqual on type paths uses set semantics (entity has no matching type in the array), which fixes multi-typed entities under exclusions. String ops on array-backed paths fail at compile time with UnsupportedTextArrayOperation. Tests add SQL snapshot coverage; friendship integration filter switches from startsWith on type.baseUrl to equal.

^{Reviewed by Cursor Bugbot for commit d6638eb. Bugbot is set up for automated code reviews on this repo. Configure here.}

[Copilot]

Pull request overview

This PR optimizes entity type filtering in the Postgres SelectCompiler by compiling type predicates into GIN-indexable array operations over entity_edition_cache (instead of generating one type-join chain per predicate), reducing join multiplication and planner misestimates.

Changes:

• Introduces an ArrayContains (@>) binary operator and compiles eligible type/base-url filters into @> / && predicates (with bundling inside All/Any groups).
• Updates query-path resolution so EntityTypeEdge { IsOfType, BaseUrl|VersionedUrl, inheritance_depth: None } maps directly to entity_edition_cache type arrays; removes redundant TypeBaseUrls/TypeVersionedUrls paths.
• Adds/updates snapshot & integration coverage to pin the new compilation output and alias-separation behavior.

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tests/graph/integration/postgres/email_filter_protection.rs	Updates protection-filter paths to the canonical EntityTypeEdge form (depth None) used by array-backed compilation.
libs/@local/graph/store/src/filter/protection.rs	Switches protection config/type-path usage to EntityTypeEdge and removes references to deleted Type*Urls paths.
libs/@local/graph/store/src/filter/mod.rs	Rewrites for_entity_by_type_id to a single versionedUrl equality filter suited to array-backed compilation.
libs/@local/graph/store/src/entity/query.rs	Removes TypeBaseUrls/TypeVersionedUrls query paths and updates docs for DirectTypeCount semantics with EntityTypeEdge.
libs/@local/graph/postgres-store/src/store/postgres/query/statement/select.rs	Adds snapshot tests covering single predicates, All/Any bundling, negated bundling, and alias-separation regression cases.
libs/@local/graph/postgres-store/src/store/postgres/query/expression/conditional.rs	Adds Expression::array_contains helper constructing BinaryOperator::ArrayContains.
libs/@local/graph/postgres-store/src/store/postgres/query/expression/binary.rs	Adds BinaryOperator::ArrayContains and transpilation support (@>).
libs/@local/graph/postgres-store/src/store/postgres/query/entity.rs	Maps EntityTypeEdge (IsOfType + BaseUrl/VersionedUrl + no depth) to entity_edition_cache relations/columns.
libs/@local/graph/postgres-store/src/store/postgres/query/compile.rs	Implements text-array predicate compilation and All/Any bundling; rejects string ops on array-backed paths at compile time.
libs/@local/graph/postgres-store/src/store/postgres/knowledge/entity/summary.rs	Updates summary selection to use EntityTypeEdge(...VersionedUrl, depth None) for type IDs/titles plumbing.
libs/@local/graph/postgres-store/src/store/postgres/knowledge/entity/query.rs	Updates entity record selection to use the new canonical type path (EntityTypeEdge to cached versioned_urls).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

Benchmark results

@rust/hash-graph-benches – Integrations

policy_resolution_large

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2002	$$27.2 \mathrm{ms} \pm 207 \mathrm{μs}\left({\color{gray}-1.757 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$3.53 \mathrm{ms} \pm 24.7 \mathrm{μs}\left({\color{gray}3.54 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 1001	$$12.2 \mathrm{ms} \pm 101 \mathrm{μs}\left({\color{gray}-1.712 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 3314	$$42.7 \mathrm{ms} \pm 479 \mathrm{μs}\left({\color{gray}-2.277 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$14.9 \mathrm{ms} \pm 135 \mathrm{μs}\left({\color{gray}2.40 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 1526	$$23.7 \mathrm{ms} \pm 245 \mathrm{μs}\left({\color{gray}-3.304 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 2078	$$27.9 \mathrm{ms} \pm 215 \mathrm{μs}\left({\color{gray}-1.907 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$3.80 \mathrm{ms} \pm 24.5 \mathrm{μs}\left({\color{gray}2.04 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 1033	$$13.6 \mathrm{ms} \pm 131 \mathrm{μs}\left({\color{gray}0.658 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_medium

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 102	$$3.99 \mathrm{ms} \pm 36.0 \mathrm{μs}\left({\color{gray}2.56 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$3.22 \mathrm{ms} \pm 35.6 \mathrm{μs}\left({\color{red}6.69 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 51	$$3.54 \mathrm{ms} \pm 28.4 \mathrm{μs}\left({\color{gray}4.71 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 269	$$5.35 \mathrm{ms} \pm 45.7 \mathrm{μs}\left({\color{gray}3.59 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$3.51 \mathrm{ms} \pm 22.5 \mathrm{μs}\left({\color{gray}-1.331 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 107	$$4.09 \mathrm{ms} \pm 23.3 \mathrm{μs}\left({\color{gray}-2.287 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 133	$$4.50 \mathrm{ms} \pm 33.8 \mathrm{μs}\left({\color{gray}-0.121 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$3.64 \mathrm{ms} \pm 26.8 \mathrm{μs}\left({\color{gray}4.45 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 63	$$4.13 \mathrm{ms} \pm 30.7 \mathrm{μs}\left({\color{gray}-0.795 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_none

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2	$$2.73 \mathrm{ms} \pm 22.8 \mathrm{μs}\left({\color{gray}2.32 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.74 \mathrm{ms} \pm 26.0 \mathrm{μs}\left({\color{red}8.44 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 1	$$2.75 \mathrm{ms} \pm 21.1 \mathrm{μs}\left({\color{red}5.35 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 8	$$3.08 \mathrm{ms} \pm 29.0 \mathrm{μs}\left({\color{red}8.16 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.78 \mathrm{ms} \pm 19.1 \mathrm{μs}\left({\color{gray}2.40 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 3	$$2.89 \mathrm{ms} \pm 22.8 \mathrm{μs}\left({\color{gray}0.974 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_small

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 52	$$3.03 \mathrm{ms} \pm 21.7 \mathrm{μs}\left({\color{gray}-1.940 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.72 \mathrm{ms} \pm 16.8 \mathrm{μs}\left({\color{gray}-2.503 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 25	$$2.96 \mathrm{ms} \pm 15.1 \mathrm{μs}\left({\color{gray}-1.631 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 94	$$3.41 \mathrm{ms} \pm 20.5 \mathrm{μs}\left({\color{gray}-1.223 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$2.95 \mathrm{ms} \pm 29.7 \mathrm{μs}\left({\color{gray}-0.829 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 26	$$3.25 \mathrm{ms} \pm 19.3 \mathrm{μs}\left({\color{gray}-1.801 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 66	$$3.32 \mathrm{ms} \pm 19.6 \mathrm{μs}\left({\color{gray}-2.035 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.90 \mathrm{ms} \pm 14.8 \mathrm{μs}\left({\color{gray}-1.372 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 29	$$3.39 \mathrm{ms} \pm 30.0 \mathrm{μs}\left({\color{gray}0.650 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_complete

Function	Value	Mean	Flame graphs
entity_by_id;one_depth	1 entities	$$42.0 \mathrm{ms} \pm 301 \mathrm{μs}\left({\color{lightgreen}-29.851 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	10 entities	$$32.4 \mathrm{ms} \pm 305 \mathrm{μs}\left({\color{lightgreen}-35.930 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	25 entities	$$34.9 \mathrm{ms} \pm 182 \mathrm{μs}\left({\color{lightgreen}-34.964 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	5 entities	$$31.1 \mathrm{ms} \pm 173 \mathrm{μs}\left({\color{lightgreen}-36.863 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	50 entities	$$40.4 \mathrm{ms} \pm 230 \mathrm{μs}\left({\color{lightgreen}-37.645 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	1 entities	$$48.6 \mathrm{ms} \pm 265 \mathrm{μs}\left({\color{lightgreen}-26.983 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	10 entities	$$39.3 \mathrm{ms} \pm 201 \mathrm{μs}\left({\color{lightgreen}-31.543 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	25 entities	$$91.7 \mathrm{ms} \pm 559 \mathrm{μs}\left({\color{lightgreen}-14.828 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	5 entities	$$33.0 \mathrm{ms} \pm 226 \mathrm{μs}\left({\color{lightgreen}-38.081 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	50 entities	$$280 \mathrm{ms} \pm 1.00 \mathrm{ms}\left({\color{lightgreen}-7.139 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	1 entities	$$10.6 \mathrm{ms} \pm 70.6 \mathrm{μs}\left({\color{lightgreen}-51.089 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	10 entities	$$10.5 \mathrm{ms} \pm 71.9 \mathrm{μs}\left({\color{lightgreen}-54.135 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	25 entities	$$10.6 \mathrm{ms} \pm 68.8 \mathrm{μs}\left({\color{lightgreen}-53.165 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	5 entities	$$10.4 \mathrm{ms} \pm 61.8 \mathrm{μs}\left({\color{lightgreen}-53.302 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	50 entities	$$10.5 \mathrm{ms} \pm 72.2 \mathrm{μs}\left({\color{lightgreen}-61.683 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_linkless

Function	Value	Mean	Flame graphs
entity_by_id	1 entities	$$10.4 \mathrm{ms} \pm 56.0 \mathrm{μs}\left({\color{lightgreen}-52.838 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10 entities	$$10.6 \mathrm{ms} \pm 66.7 \mathrm{μs}\left({\color{lightgreen}-51.672 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	100 entities	$$10.5 \mathrm{ms} \pm 59.9 \mathrm{μs}\left({\color{lightgreen}-52.238 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	1000 entities	$$10.4 \mathrm{ms} \pm 60.6 \mathrm{μs}\left({\color{lightgreen}-51.909 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10000 entities	$$10.6 \mathrm{ms} \pm 62.8 \mathrm{μs}\left({\color{lightgreen}-54.320 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity

Function	Value	Mean	Flame graphs
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/block/v/1	$$11.0 \mathrm{ms} \pm 71.8 \mathrm{μs}\left({\color{lightgreen}-59.286 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/book/v/1	$$10.8 \mathrm{ms} \pm 55.4 \mathrm{μs}\left({\color{lightgreen}-56.752 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/building/v/1	$$10.8 \mathrm{ms} \pm 80.6 \mathrm{μs}\left({\color{lightgreen}-57.250 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/organization/v/1	$$10.8 \mathrm{ms} \pm 67.4 \mathrm{μs}\left({\color{lightgreen}-56.502 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/page/v/2	$$10.8 \mathrm{ms} \pm 65.1 \mathrm{μs}\left({\color{lightgreen}-56.592 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/person/v/1	$$10.8 \mathrm{ms} \pm 60.7 \mathrm{μs}\left({\color{lightgreen}-57.466 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/playlist/v/1	$$11.1 \mathrm{ms} \pm 92.0 \mathrm{μs}\left({\color{lightgreen}-55.474 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/song/v/1	$$10.8 \mathrm{ms} \pm 57.8 \mathrm{μs}\left({\color{lightgreen}-57.977 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: https://blockprotocol.org/@alice/types/entity-type/uk-address/v/1	$$10.8 \mathrm{ms} \pm 67.2 \mathrm{μs}\left({\color{lightgreen}-60.322 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity_type

Function	Value	Mean	Flame graphs
get_entity_type_by_id	Account ID: bf5a9ef5-dc3b-43cf-a291-6210c0321eba	$$8.65 \mathrm{ms} \pm 68.5 \mathrm{μs}\left({\color{gray}-1.992 \mathrm{\%}}\right) $$	Flame Graph

representative_read_multiple_entities

Function	Value	Mean	Flame graphs
entity_by_property	traversal_paths=0	0	$$57.9 \mathrm{ms} \pm 395 \mathrm{μs}\left({\color{lightgreen}-8.961 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$109 \mathrm{ms} \pm 610 \mathrm{μs}\left({\color{lightgreen}-7.480 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$62.9 \mathrm{ms} \pm 417 \mathrm{μs}\left({\color{lightgreen}-10.047 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$73.1 \mathrm{ms} \pm 541 \mathrm{μs}\left({\color{lightgreen}-9.124 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$81.6 \mathrm{ms} \pm 501 \mathrm{μs}\left({\color{lightgreen}-11.186 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$88.6 \mathrm{ms} \pm 601 \mathrm{μs}\left({\color{lightgreen}-7.248 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=0	0	$$43.1 \mathrm{ms} \pm 272 \mathrm{μs}\left({\color{lightgreen}-47.422 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$72.5 \mathrm{ms} \pm 429 \mathrm{μs}\left({\color{lightgreen}-35.119 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$49.8 \mathrm{ms} \pm 297 \mathrm{μs}\left({\color{lightgreen}-45.140 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$60.0 \mathrm{ms} \pm 523 \mathrm{μs}\left({\color{lightgreen}-39.114 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$61.6 \mathrm{ms} \pm 397 \mathrm{μs}\left({\color{lightgreen}-38.330 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$61.7 \mathrm{ms} \pm 426 \mathrm{μs}\left({\color{lightgreen}-37.960 \mathrm{\%}}\right) $$

scenarios

Function	Value	Mean	Flame graphs
full_test	query-limited	$$118 \mathrm{ms} \pm 702 \mathrm{μs}\left({\color{lightgreen}-16.011 \mathrm{\%}}\right) $$	Flame Graph
full_test	query-unlimited	$$129 \mathrm{ms} \pm 692 \mathrm{μs}\left({\color{lightgreen}-14.286 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-limited	$$23.6 \mathrm{ms} \pm 191 \mathrm{μs}\left({\color{lightgreen}-45.578 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-unlimited	$$548 \mathrm{ms} \pm 1.42 \mathrm{ms}\left({\color{gray}-2.750 \mathrm{\%}}\right) $$	Flame Graph