[Backport stable-26-3] PR #49216: Reduce test size under sanitizer

[ydbot]

Changelog entry

Reduce test size under sanitizer

Description for reviewers

Original PR(s)

• PR Reduce test size under sanitizer #49216

Metadata

• Original PR author(s): @zverevgeny
• Cherry-picked by: @ydbot
• Related issues: [sanitizer] KqpOlapScheme.TenThousandColumns VERIFY / BulkUpsert TIMEOUT in cs_helper.cpp:87 (release-tsan) #48623

Git Cherry-Pick Log

=== Cherry-picking 8ed5b62 ===
Auto-merging ydb/core/kqp/ut/scheme/kqp_scheme_ut.cpp
[cherry-pick-stable-26-3-260807-084907 0dc660c2ccf] Reduce test size under sanitizer (#49216)
 Author: zverevgeny <zverevgeny@ydb.tech>
 Date: Fri Aug 7 11:39:57 2026 +0300
 1 file changed, 17 insertions(+), 5 deletions(-)

---

PR was created by cherry-pick workflow run

[ydbot]

Run Extra Tests

Run additional tests for this PR. You can customize:

• Test Size: small, medium, large (default: small, medium)
• Test Targets: any directory path (default: ydb/)
• Sanitizers: ASAN, MSAN, TSAN
• Coredumps: enable for debugging (default: off)
• Additional args: custom ya make arguments

[github-actions]

⚪ 2026-08-07 08:50:54 UTC Pre-commit check linux-x86_64-relwithdebinfo for 2753ec4 has started.
⚪ 2026-08-07 08:52:18 UTC Artifacts will be uploaded here
⚪ 2026-08-07 08:53:53 UTC ya make is running...
🟢 2026-08-07 08:58:13 UTC Tests successful.

Ya make output | Test bloat

TESTS	PASSED	ERRORS	FAILED	SKIPPED	MUTED?
606	606	0	0	0	0

🟢 2026-08-07 08:58:16 UTC Build successful.

[github-actions]

⚪ 2026-08-07 08:50:54 UTC Pre-commit check linux-x86_64-release-asan for 2753ec4 has started.
⚪ 2026-08-07 08:53:34 UTC Artifacts will be uploaded here
⚪ 2026-08-07 08:55:04 UTC ya make is running...
🟡 2026-08-07 09:04:18 UTC Some tests failed, follow the links below. Going to retry failed tests...

Details

Ya make output | Test bloat

TESTS	PASSED	ERRORS	FAILED	SKIPPED	MUTED?
606	605	0	1	0	0

⚪ 2026-08-07 09:04:36 UTC ya make is running... (failed tests rerun, try 2)
🟢 2026-08-07 09:05:20 UTC Tests successful.

Ya make output | Test bloat | Test bloat

TESTS	PASSED	ERRORS	FAILED	SKIPPED	MUTED?
51 (only retried tests)	51	0	0	0	0

🟢 2026-08-07 09:05:23 UTC Build successful.

[robot-vibe-db]

AI Review Summary

Verdict: No critical issues found

Critical issues

No critical issues found.

Other findings

• Nit | Low: numColumns, alterColumnsFrom, alterColumnsTo are constant-valued (no PlainOrUnderSanitizer) -- they serve as named constants replacing magic numbers, which is a good readability improvement but not strictly necessary for the sanitizer fix. No action needed. -- ydb/core/kqp/ut/scheme/kqp_scheme_ut.cpp:16531-16534

Details

This is a clean backport cherry-pick of PR #49216. The diff is identical to the original (only line numbers differ due to the stable-26-3 branch). The change is correct and follows established patterns in the codebase:

1. NSan::PlainOrUnderSanitizer usage matches the well-established pattern used in many other test files (btree_cow_ut.cpp, kqp_scripting_ut.cpp, kqp_service_ut.cpp, etc.).

2. Row count of 100 under sanitizers is reasonable -- many other tests in the same file already use BuildBatch(100) as their standard batch size, so this does not undermine test coverage.

3. Column count is preserved at 9900 (not reduced under sanitizers), which correctly preserves the core purpose of the test: validating wide-schema behavior with approximately 10,000 columns.

4. Expected result strings are correctly made dynamic via TStringBuilder() << "[[" << numRows << "u]]", which produces "[[10000u]]" in plain builds and "[[100u]]" under sanitizers. The ui64 type streams correctly as a decimal string.

5. No magic number mismatches: all four original hardcoded values (9900, 10000, 9900, 9999) are faithfully captured in the named constants.

6. The comment accurately states "9901-column": 9900 data columns + 1 "id" primary key column = 9901 total columns.

No correctness, crash, data-loss, security, concurrency, or performance issues found.

---

This review was generated automatically. Critical issues require attention; other findings are advisory.
If this comment was useful, please give it a 👍 -- it helps us improve the review bot.

[robot-vibe-db]

Severity: Nit
Confidence: Low

numColumns, alterColumnsFrom, and alterColumnsTo are constant-valued -- they don't use PlainOrUnderSanitizer and always evaluate to 9900, 9900, and 9999 respectively. They serve purely as named constants replacing magic numbers.

This is a nice readability improvement (makes the relationship between the values explicit and easier to change in the future), but it goes slightly beyond the minimal sanitizer fix. No action needed -- just noting for completeness.