YTDB-630: Skip path construction

[sandrawar]

PR Title:

YTDB-630: skip path construction

Motivation:

The IS2 query does not declare a pathAlias. The pathToHere parameter is only needed when the user requests $matchPath metadata. When no pathAlias exists, pass null and skip all path construction

Expected impact: For IS2, this eliminates all path ArrayList allocations and copies. Combined with R4a, this is the most impactful allocation reduction.

Risk: Low — controlled by a simple null check. The pathToHere is already nullable.

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces a significant performance optimization by conditionally skipping the construction of path-related data structures and metadata when a pathAlias is not specified in a query. This change primarily targets queries like IS2, which do not declare a pathAlias, leading to a substantial reduction in ArrayList allocations and copies. The modification is controlled by a simple null check, ensuring low risk while delivering high impact on memory efficiency.

Highlights

• Conditional PathNode Construction: PathNode objects are now only constructed if a pathAlias is explicitly declared, preventing unnecessary allocations for queries that do not require path metadata.
• Conditional $matchPath Metadata Setting: The $matchPath metadata is now only set if a pathAlias is present, further reducing overhead when path information is not requested.
• Refactored Path Alias Check: Introduced a hasPathAlias boolean variable to centralize the check for pathAlias existence, improving readability and consistency in the MatchEdgeTraverser.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces performance optimizations to the MatchEdgeTraverser by conditionally creating path-related objects and metadata only when a path alias is explicitly declared in the query. This change aims to reduce memory allocation and processing for queries that do not require path information. A new test case, testWhileWithoutPathAliasSkipsPathConstruction, has been added to validate this optimization, ensuring that traversals without path aliases still correctly identify reachable vertices and depths while confirming that path construction is skipped. There is no feedback to provide.

[github-actions]

Coverage Gate Results

Thresholds: 85% line, 70% branch

Line Coverage: ✅ 100.0% (6/6 lines)

File	Coverage	Uncovered Lines
core/src/main/java/com/jetbrains/youtrackdb/internal/core/sql/executor/match/MatchEdgeTraverser.java	✅ 100.0% (6/6)	-

Branch Coverage: ✅ 100.0% (10/10 branches)

File	Coverage	Lines with Uncovered Branches
core/src/main/java/com/jetbrains/youtrackdb/internal/core/sql/executor/match/MatchEdgeTraverser.java	✅ 100.0% (10/10)	-

[github-actions]

Test Count Gate Results

✅ No baseline available yet — gate skipped (first run).

[github-actions]

JMH LDBC Benchmark Comparison

Base: 244d8c7275 (fork-point with develop) | Head: f1f81cc418
Summary: 🟢 28 improvement(s) (>±5% threshold)

Single-Thread Results

Benchmark	Base ops/s	Base err	Head ops/s	Head err	Δ%
ic10_friendRecommendation	0.137	±5.2%	0.145	±5.9%	+6.1% 🟢
ic11_jobReferral	36.0	±5.8%	40.1	±1.6%	+11.3% 🟢
ic12_expertSearch	22.2	±5.6%	25.0	±0.6%	+12.4% 🟢
ic13_shortestPath	4,169	±1.3%	4,471	±1.5%	+7.2% 🟢
ic1_transitiveFriends	39.4	±5.9%	46.1	±0.9%	+16.8% 🟢
ic2_recentFriendMessages	205.2	±9.5%	237.7	±1.5%	+15.8% 🟢
ic3_friendsInCountries	0.162	±5.5%	0.177	±5.6%	+9.5% 🟢
ic4_newTopics	2.7	±6.9%	3.4	±1.7%	+27.6% 🟢
ic5_newGroups	0.095	±24.2%	0.096	±24.1%	+1.2%
ic6_tagCoOccurrence	3.7	±5.1%	3.9	±2.9%	+5.2% 🟢
ic7_recentLikers	56.4	±4.9%	64.2	±2.5%	+14.0% 🟢
ic8_recentReplies	971.0	±0.8%	971.2	±0.5%	+0.0%
ic9_recentFofMessages	1.3	±4.6%	1.3	±1.6%	+3.9%
is1_personProfile	56,631	±1.0%	57,261	±1.5%	+1.1%
is2_personPosts	566.4	±1.1%	580.8	±1.1%	+2.5%
is3_personFriends	16,328	±3.0%	16,581	±3.2%	+1.6%
is4_messageContent	79,602	±0.6%	79,540	±0.8%	-0.1%
is5_messageCreator	71,704	±0.8%	71,322	±2.3%	-0.5%
is6_messageForum	49,362	±1.6%	48,940	±1.7%	-0.9%
is7_messageReplies	2,870	±0.6%	3,466	±0.6%	+20.7% 🟢

Multi-Thread Results

Benchmark	Base ops/s	Base err	Head ops/s	Head err	Δ%
ic10_friendRecommendation	0.589	±1.8%	0.616	±3.3%	+4.6%
ic11_jobReferral	160.5	±3.5%	177.5	±3.2%	+10.6% 🟢
ic12_expertSearch	113.6	±2.0%	126.6	±3.0%	+11.4% 🟢
ic13_shortestPath	16,369	±4.0%	19,724	±3.8%	+20.5% 🟢
ic1_transitiveFriends	185.9	±3.5%	211.1	±1.9%	+13.6% 🟢
ic2_recentFriendMessages	882.8	±2.2%	1,036	±2.6%	+17.4% 🟢
ic3_friendsInCountries	0.630	±7.4%	0.704	±2.2%	+11.7% 🟢
ic4_newTopics	9.3	±6.9%	13.8	±3.2%	+48.4% 🟢
ic5_newGroups	0.331	±8.3%	0.409	±7.8%	+23.7% 🟢
ic6_tagCoOccurrence	15.7	±2.1%	17.7	±2.6%	+12.8% 🟢
ic7_recentLikers	241.7	±6.9%	277.2	±2.5%	+14.7% 🟢
ic8_recentReplies	4,342	±1.2%	4,730	±1.7%	+8.9% 🟢
ic9_recentFofMessages	5.8	±4.8%	6.3	±3.6%	+8.6% 🟢
is1_personProfile	223,078	±2.6%	232,992	±2.4%	+4.4%
is2_personPosts	2,652	±1.2%	2,844	±2.6%	+7.3% 🟢
is3_personFriends	66,613	±3.8%	71,147	±4.2%	+6.8% 🟢
is4_messageContent	291,725	±2.0%	312,971	±1.5%	+7.3% 🟢
is5_messageCreator	267,660	±1.8%	287,262	±1.0%	+7.3% 🟢
is6_messageForum	190,005	±1.9%	199,160	±5.1%	+4.8%
is7_messageReplies	11,735	±2.1%	14,931	±2.7%	+27.2% 🟢

Scalability (MT/ST ratio)

Benchmark	Base ratio	Head ratio	Δ%
ic10_friendRecommendation	4.32x	4.25x	-1.4%
ic11_jobReferral	4.46x	4.43x	-0.7%
ic12_expertSearch	5.11x	5.06x	-1.0%
ic13_shortestPath	3.93x	4.41x	+12.4%
ic1_transitiveFriends	4.71x	4.58x	-2.8%
ic2_recentFriendMessages	4.30x	4.36x	+1.4%
ic3_friendsInCountries	3.90x	3.97x	+2.0%
ic4_newTopics	3.47x	4.03x	+16.3%
ic5_newGroups	3.50x	4.27x	+22.2%
ic6_tagCoOccurrence	4.28x	4.59x	+7.3%
ic7_recentLikers	4.29x	4.32x	+0.6%
ic8_recentReplies	4.47x	4.87x	+8.9%
ic9_recentFofMessages	4.57x	4.78x	+4.6%
is1_personProfile	3.94x	4.07x	+3.3%
is2_personPosts	4.68x	4.90x	+4.6%
is3_personFriends	4.08x	4.29x	+5.2%
is4_messageContent	3.66x	3.93x	+7.4%
is5_messageCreator	3.73x	4.03x	+7.9%
is6_messageForum	3.85x	4.07x	+5.7%
is7_messageReplies	4.09x	4.31x	+5.4%

[andrii0lomakin]

Hi @sandrawar, please profile regressions using asyncprofiler on Hetzner CCX 33 node and find out what caused the regressions. It should be quite the opposite as I see it.