BE-702: Compile sorted entity reads as keys-first statements by TimDiekmann · Pull Request #9048 · hashintel/hash

TimDiekmann · 2026-07-17T13:02:04Z

🌟 What is the purpose of this PR?

Sorted and limited entity reads now compile as a keys-first statement: a narrow key query carries the filters, cursor, and sort, and the wide projection is hydrated only for the rows that survive the LIMIT. This keeps the planner's ability to serve the ordering from an index while it sorts and spools far narrower rows, and it is the structural groundwork for optimizing the entity read path against known query shapes.

Grounded in the created-at sort incident: the pathological plan was reproduced locally against a 700k-entity seed and the shape trade-offs were measured with result-set-equivalence checks on every variant (see the Slack thread for the discussion).

Stacked on #9033; only the commits of this branch are new.

🔗 Related links

BE-702 (internal) — the driving ticket
BE-598 (internal) — the investigation this implements
Slack: entities-page performance & sort keys (internal) — incident context, measurements, and the follow-up direction
Slack: entities table loading behaviour (internal)

🚫 Blocked by

BE-641: Extend the SQL AST along gram.y and make invalid SQL unrepresentable #9033

🔍 What does this change?

StatementShape on the SelectCompiler: SinglePass (default, unchanged) | KeysFirst | FencedKeysFirst. The keys-first shapes compile into a key CTE (filters + cursor + sort columns), a page CTE (ORDER BY + LIMIT), and a hydration SELECT with all key-table references rewritten; the fenced variant adds AS MATERIALIZED to shield the key query from the outer ordering demand.
Join classification by use (filters/cursor/sort keys, closed transitively over the join conditions) splits the collected joins into the key query's and the hydration's share; the FROM tree is now assembled in compile() from that bookkeeping.
Statements that cannot be split degrade to SinglePass with a logged ShapeFallback reason; the effective shape is recorded on the compile span as statement.shape.
Pre-order Expression::visit/visit_mut walkers on the AST, used by the classification and reference rewriting.
read_entities_impl opts into KeysFirst for sorted entity reads; embedding-distance queries stay on SinglePass (TODO(BE-618)).
A note on AsClient for Object documents that bypassing the pool's statement cache keeps every query on a custom plan — the shape strategy relies on parameter-aware planning.

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

FencedKeysFirst has no production caller yet; it is exercised by the compile tests and reserved as an escalation shape.
Statement texts for sorted entity reads change shape, so pg_stat_statements fingerprints reset for those queries.

🐾 Next steps

Distinguish include- from exclude-type filters on the entity read path: inclusion lists are GIN-indexable and estimable, and the entities page can derive them from the type summary it already fetches (see the Slack thread).

🛡 What tests cover this?

Golden compile tests for both keys-first shapes: sort keys across joins, cursor continuation inside the key CTE, embedding-distance queries, carried CTEs, and filter-joined tables staying out of the key projection
Fallback tests asserting the specific ShapeFallback reason plus byte-equality with the single-pass output
The full hash-graph-integration postgres suite (including all sorting/pagination tests) against a migrated database with KeysFirst active
Result-set-equivalence checks (row-set fingerprints) of every measured variant against a 700k-entity seed

❓ How to test this?

cargo nextest run --package hash-graph-postgres-store --lib
With a migrated local database: cargo nextest run --package hash-graph-integration --test postgres

📹 Demo

WITH "roots" AS (SELECT "entity_ids_0_1_0"."created_at_transaction_time", "entity_temporal_metadata_0_0_0"."entity_uuid", "entity_temporal_metadata_0_0_0"."entity_edition_id"
FROM "entity_temporal_metadata" AS "entity_temporal_metadata_0_0_0"
INNER JOIN "entity_ids" AS "entity_ids_0_1_0" ON ...
WHERE ...),
"limited" AS (SELECT * FROM "roots" ORDER BY "created_at_transaction_time" ASC LIMIT 500)
SELECT DISTINCT ON("limited"."created_at_transaction_time", ...) ..., "entity_editions_1_1_0"."properties"
FROM "limited"
INNER JOIN "entity_editions" AS "entity_editions_1_1_0" ON "entity_editions_1_1_0"."entity_edition_id" = "limited"."entity_edition_id"
ORDER BY "limited"."created_at_transaction_time" ASC LIMIT 500

…pass

vercel · 2026-07-17T13:02:28Z

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

Project	Deployment	Actions	Updated (UTC)
hash	Ready	Preview, Comment	Jul 17, 2026 2:29pm

2 Skipped Deployments

Project	Deployment	Actions	Updated (UTC)
hashdotdesign-tokens	Ignored	Preview	Jul 17, 2026 2:29pm
petrinaut	Skipped		Jul 17, 2026 2:29pm

codecov · 2026-07-17T13:11:08Z

Codecov Report

❌ Patch coverage is 73.06338% with 153 lines in your changes missing coverage. Please review.
✅ Project coverage is 59.37%. Comparing base (6753875) to head (c2b189e).

Files with missing lines	Patch %	Lines
...rc/store/postgres/query/ast/expression/function.rs	0.00%	72 Missing ⚠️
...ore/src/store/postgres/query/ast/expression/mod.rs	32.97%	63 Missing ⚠️
...gres-store/src/store/postgres/query/compile/mod.rs	95.52%	10 Missing and 8 partials ⚠️

Additional details and impacted files

@@                       Coverage Diff                       @@
##           t/be-641-extend-the-sql-ast    #9048      +/-   ##
===============================================================
+ Coverage                        59.31%   59.37%   +0.06%     
===============================================================
  Files                             1393     1393              
  Lines                           135212   135737     +525     
  Branches                          6183     6205      +22     
===============================================================
+ Hits                             80196    80590     +394     
- Misses                           54086    54212     +126     
- Partials                           930      935       +5

Flag	Coverage Δ
apps.hash-ai-worker-ts	`1.92% <ø> (ø)`
apps.hash-api	`7.59% <ø> (ø)`
local.hash-backend-utils	`1.88% <ø> (ø)`
local.hash-graph-sdk	`10.02% <ø> (ø)`
local.hash-isomorphic-utils	`0.17% <ø> (ø)`
rust.hash-graph-api	`7.41% <ø> (ø)`
rust.hash-graph-postgres-store	`27.40% <73.06%> (+1.36%)`	⬆️
rust.hashql-compiletest	`28.39% <ø> (ø)`
rust.hashql-eval	`79.82% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

cursor · 2026-07-17T14:16:29Z

PR Summary

Medium Risk
Changes core entity list SQL shape and pagination semantics if keys-first preconditions are violated; fallbacks mitigate but wrong pages are possible in edge cases. Embedding searches stay on single-pass.

Overview
Sorted, limited entity reads can now compile as a keys-first plan instead of one wide SELECT: narrow roots / limited CTEs apply filters, cursor, sort, and LIMIT, then a hydration SELECT joins back only for the page (optional MATERIALIZED via FencedKeysFirst).

SelectCompiler gains StatementShape, join bookkeeping (CompiledJoin with join type, subquery sources, per-join to_many), classification of key vs hydration joins, KeyColumnRewriter, and safe fallback to SinglePass with logged ShapeFallback reasons; effective shape is traced as statement.shape. Expression AST adds visit / visit_mut for that analysis.

Production: read_entities_impl opts into KeysFirst when there is no embeddings distance filter. Pool: docs note that AsClient for Object deref bypasses statement caching so Postgres can plan with real parameters.

^{Reviewed by Cursor Bugbot for commit c2b189e. Bugbot is set up for automated code reviews on this repo. Configure here.}

Copilot

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

github-actions · 2026-07-17T15:41:31Z

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	$$21.4 \mathrm{ms} \pm 112 \mathrm{μs}\left({\color{gray}-1.476 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.49 \mathrm{ms} \pm 11.1 \mathrm{μs}\left({\color{gray}-0.655 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 1002	$$9.74 \mathrm{ms} \pm 68.7 \mathrm{μs}\left({\color{gray}-0.693 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 3314	$$30.6 \mathrm{ms} \pm 212 \mathrm{μs}\left({\color{lightgreen}-5.892 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$9.63 \mathrm{ms} \pm 69.4 \mathrm{μs}\left({\color{gray}-2.567 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 1527	$$17.5 \mathrm{ms} \pm 118 \mathrm{μs}\left({\color{gray}-0.228 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 2078	$$22.3 \mathrm{ms} \pm 148 \mathrm{μs}\left({\color{gray}0.630 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.75 \mathrm{ms} \pm 12.7 \mathrm{μs}\left({\color{gray}-0.118 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 1033	$$10.6 \mathrm{ms} \pm 73.5 \mathrm{μs}\left({\color{gray}0.256 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_medium

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 102	$$2.83 \mathrm{ms} \pm 18.4 \mathrm{μs}\left({\color{gray}-4.007 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.22 \mathrm{ms} \pm 12.0 \mathrm{μs}\left({\color{gray}-0.746 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 52	$$2.51 \mathrm{ms} \pm 15.1 \mathrm{μs}\left({\color{gray}0.690 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 269	$$3.97 \mathrm{ms} \pm 24.4 \mathrm{μs}\left({\color{gray}-0.745 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$2.62 \mathrm{ms} \pm 15.7 \mathrm{μs}\left({\color{gray}-1.444 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 108	$$3.13 \mathrm{ms} \pm 23.0 \mathrm{μs}\left({\color{gray}0.309 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 133	$$3.37 \mathrm{ms} \pm 23.1 \mathrm{μs}\left({\color{gray}-1.305 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.55 \mathrm{ms} \pm 16.9 \mathrm{μs}\left({\color{gray}-0.063 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 63	$$2.98 \mathrm{ms} \pm 14.8 \mathrm{μs}\left({\color{gray}-3.135 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_none

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2	$$1.99 \mathrm{ms} \pm 9.99 \mathrm{μs}\left({\color{gray}-1.295 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.00 \mathrm{ms} \pm 15.3 \mathrm{μs}\left({\color{gray}1.50 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 2	$$2.04 \mathrm{ms} \pm 12.0 \mathrm{μs}\left({\color{gray}0.135 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 8	$$2.24 \mathrm{ms} \pm 12.2 \mathrm{μs}\left({\color{gray}-2.262 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.11 \mathrm{ms} \pm 11.8 \mathrm{μs}\left({\color{gray}0.908 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 3	$$2.33 \mathrm{ms} \pm 21.0 \mathrm{μs}\left({\color{gray}1.82 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_small

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 52	$$2.23 \mathrm{ms} \pm 11.2 \mathrm{μs}\left({\color{lightgreen}-6.951 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.04 \mathrm{ms} \pm 11.8 \mathrm{μs}\left({\color{gray}-3.210 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 26	$$2.14 \mathrm{ms} \pm 12.5 \mathrm{μs}\left({\color{gray}-3.713 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 94	$$2.55 \mathrm{ms} \pm 13.1 \mathrm{μs}\left({\color{lightgreen}-6.319 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$2.24 \mathrm{ms} \pm 12.0 \mathrm{μs}\left({\color{gray}-1.807 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 27	$$2.46 \mathrm{ms} \pm 12.8 \mathrm{μs}\left({\color{gray}-2.011 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 66	$$2.51 \mathrm{ms} \pm 15.8 \mathrm{μs}\left({\color{gray}-4.023 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.22 \mathrm{ms} \pm 11.2 \mathrm{μs}\left({\color{lightgreen}-6.877 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 29	$$2.44 \mathrm{ms} \pm 13.9 \mathrm{μs}\left({\color{gray}-4.653 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_complete

Function	Value	Mean	Flame graphs
entity_by_id;one_depth	1 entities	$$25.5 \mathrm{ms} \pm 130 \mathrm{μs}\left({\color{lightgreen}-5.412 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	10 entities	$$55.7 \mathrm{ms} \pm 319 \mathrm{μs}\left({\color{gray}-2.485 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	25 entities	$$27.7 \mathrm{ms} \pm 135 \mathrm{μs}\left({\color{gray}-2.317 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	5 entities	$$31.7 \mathrm{ms} \pm 154 \mathrm{μs}\left({\color{gray}-3.590 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	50 entities	$$32.8 \mathrm{ms} \pm 189 \mathrm{μs}\left({\color{gray}-1.357 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	1 entities	$$28.0 \mathrm{ms} \pm 152 \mathrm{μs}\left({\color{gray}0.893 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	10 entities	$$340 \mathrm{ms} \pm 827 \mathrm{μs}\left({\color{gray}-1.161 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	25 entities	$$71.4 \mathrm{ms} \pm 466 \mathrm{μs}\left({\color{gray}-1.910 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	5 entities	$$69.1 \mathrm{ms} \pm 387 \mathrm{μs}\left({\color{gray}-1.880 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	50 entities	$$231 \mathrm{ms} \pm 1.26 \mathrm{ms}\left({\color{red}6.68 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	1 entities	$$8.80 \mathrm{ms} \pm 49.5 \mathrm{μs}\left({\color{gray}1.18 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	10 entities	$$8.73 \mathrm{ms} \pm 46.5 \mathrm{μs}\left({\color{gray}-0.711 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	25 entities	$$8.83 \mathrm{ms} \pm 41.6 \mathrm{μs}\left({\color{gray}0.870 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	5 entities	$$8.81 \mathrm{ms} \pm 41.4 \mathrm{μs}\left({\color{gray}-2.104 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	50 entities	$$8.81 \mathrm{ms} \pm 45.1 \mathrm{μs}\left({\color{gray}-0.746 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_linkless

Function	Value	Mean	Flame graphs
entity_by_id	1 entities	$$8.69 \mathrm{ms} \pm 52.2 \mathrm{μs}\left({\color{gray}0.860 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10 entities	$$8.75 \mathrm{ms} \pm 45.1 \mathrm{μs}\left({\color{gray}2.13 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	100 entities	$$8.68 \mathrm{ms} \pm 54.9 \mathrm{μs}\left({\color{gray}0.535 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	1000 entities	$$8.81 \mathrm{ms} \pm 40.7 \mathrm{μs}\left({\color{gray}0.553 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10000 entities	$$8.98 \mathrm{ms} \pm 48.4 \mathrm{μs}\left({\color{gray}0.125 \mathrm{\%}}\right) $$	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`	$$9.09 \mathrm{ms} \pm 60.3 \mathrm{μs}\left({\color{gray}0.316 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/book/v/1`	$$9.12 \mathrm{ms} \pm 42.7 \mathrm{μs}\left({\color{gray}0.840 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/building/v/1`	$$8.96 \mathrm{ms} \pm 48.5 \mathrm{μs}\left({\color{gray}-1.012 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/organization/v/1`	$$9.09 \mathrm{ms} \pm 57.1 \mathrm{μs}\left({\color{gray}1.93 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/page/v/2`	$$9.19 \mathrm{ms} \pm 51.7 \mathrm{μs}\left({\color{gray}3.16 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/person/v/1`	$$9.15 \mathrm{ms} \pm 47.4 \mathrm{μs}\left({\color{gray}2.18 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/playlist/v/1`	$$8.96 \mathrm{ms} \pm 50.0 \mathrm{μs}\left({\color{gray}0.480 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/song/v/1`	$$9.16 \mathrm{ms} \pm 52.0 \mathrm{μs}\left({\color{gray}0.581 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/uk-address/v/1`	$$9.10 \mathrm{ms} \pm 58.1 \mathrm{μs}\left({\color{gray}0.281 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity_type

Function	Value	Mean	Flame graphs
get_entity_type_by_id	Account ID: `bf5a9ef5-dc3b-43cf-a291-6210c0321eba`	$$6.52 \mathrm{ms} \pm 30.9 \mathrm{μs}\left({\color{gray}1.32 \mathrm{\%}}\right) $$	Flame Graph

representative_read_multiple_entities

Function	Value	Mean	Flame graphs
entity_by_property	traversal_paths=0	0	$$48.6 \mathrm{ms} \pm 271 \mathrm{μs}\left({\color{gray}-1.880 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$93.9 \mathrm{ms} \pm 445 \mathrm{μs}\left({\color{gray}0.653 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$54.2 \mathrm{ms} \pm 330 \mathrm{μs}\left({\color{lightgreen}-5.251 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$62.5 \mathrm{ms} \pm 381 \mathrm{μs}\left({\color{gray}-3.262 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$71.1 \mathrm{ms} \pm 400 \mathrm{μs}\left({\color{gray}-2.294 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$77.2 \mathrm{ms} \pm 397 \mathrm{μs}\left({\color{gray}1.93 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=0	0	$$38.6 \mathrm{ms} \pm 214 \mathrm{μs}\left({\color{gray}1.30 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$58.3 \mathrm{ms} \pm 311 \mathrm{μs}\left({\color{gray}0.108 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$42.8 \mathrm{ms} \pm 254 \mathrm{μs}\left({\color{gray}0.724 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$50.7 \mathrm{ms} \pm 413 \mathrm{μs}\left({\color{gray}4.09 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$50.7 \mathrm{ms} \pm 277 \mathrm{μs}\left({\color{gray}1.09 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$50.2 \mathrm{ms} \pm 338 \mathrm{μs}\left({\color{gray}0.241 \mathrm{\%}}\right) $$

scenarios

Function	Value	Mean	Flame graphs
full_test	query-limited	$$89.2 \mathrm{ms} \pm 481 \mathrm{μs}\left({\color{lightgreen}-8.995 \mathrm{\%}}\right) $$	Flame Graph
full_test	query-unlimited	$$99.3 \mathrm{ms} \pm 514 \mathrm{μs}\left({\color{lightgreen}-6.877 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-limited	$$18.2 \mathrm{ms} \pm 94.2 \mathrm{μs}\left({\color{gray}-4.182 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-unlimited	$$417 \mathrm{ms} \pm 1.01 \mathrm{ms}\left({\color{gray}-2.128 \mathrm{\%}}\right) $$	Flame Graph

TimDiekmann added 5 commits July 16, 2026 14:25

BE-702: Add pre-order expression walkers to the AST

b3774b6

BE-702: Compile sorted and limited statements as fetch-keys-then-hydrate

0aaacc5

BE-702: Opt the entity read path into fetch-keys-then-hydrate

d509c6a

BE-702: Split the keys-first shape from its materialized fence

fed5c8f

BE-702: Harden the keys-first fallbacks and pin embeddings to single-…

bad3a88

…pass

github-actions Bot added area/libs Relates to first-party libraries/crates/packages (area) type/eng > backend Owned by the @backend team labels Jul 17, 2026

vercel Bot deployed to Preview – hash July 17, 2026 13:18 View deployment

Remove link to fallback reason

c2b189e

vercel Bot temporarily deployed to Preview – petrinaut July 17, 2026 14:16 Inactive

TimDiekmann marked this pull request as ready for review July 17, 2026 14:16

graphite-app Bot assigned TimDiekmann Jul 17, 2026

TimDiekmann requested a review from Copilot July 17, 2026 14:18

Copilot started reviewing on behalf of TimDiekmann July 17, 2026 14:18 View session

Copilot AI reviewed Jul 17, 2026

vercel Bot deployed to Preview – hash July 17, 2026 14:29 View deployment

TimDiekmann requested a review from a team July 17, 2026 14:37

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

BE-702: Compile sorted entity reads as keys-first statements#9048

BE-702: Compile sorted entity reads as keys-first statements#9048
TimDiekmann wants to merge 6 commits into
t/be-641-extend-the-sql-astfrom
t/be-598-investigate-fetch-keys-then-hydrate-projection-for-the

TimDiekmann commented Jul 17, 2026

Uh oh!

vercel Bot commented Jul 17, 2026 •

edited

Loading

Uh oh!

codecov Bot commented Jul 17, 2026 •

edited

Loading

Uh oh!

cursor Bot commented Jul 17, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

github-actions Bot commented Jul 17, 2026

@rust/hash-graph-benches – Integrations

policy_resolution_large

policy_resolution_medium

policy_resolution_none

policy_resolution_small

read_scaling_complete

read_scaling_linkless

representative_read_entity

representative_read_entity_type

representative_read_multiple_entities

scenarios

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

TimDiekmann commented Jul 17, 2026

🌟 What is the purpose of this PR?

🔗 Related links

🚫 Blocked by

🔍 What does this change?

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

📜 Does this require a change to the docs?

🕸️ Does this require a change to the Turbo Graph?

⚠️ Known issues

🐾 Next steps

🛡 What tests cover this?

❓ How to test this?

📹 Demo

Uh oh!

vercel Bot commented Jul 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov Bot commented Jul 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

cursor Bot commented Jul 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Jul 17, 2026

Benchmark results

@rust/hash-graph-benches – Integrations

policy_resolution_large

policy_resolution_medium

policy_resolution_none

policy_resolution_small

read_scaling_complete

read_scaling_linkless

representative_read_entity

representative_read_entity_type

representative_read_multiple_entities

scenarios

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

2 participants

vercel Bot commented Jul 17, 2026 •

edited

Loading

codecov Bot commented Jul 17, 2026 •

edited

Loading

cursor Bot commented Jul 17, 2026 •

edited

Loading