The lance-bench run for Lance SHA 6484a0b flagged high/medium-confidence regressions in the Rust IO \ benchmarks.\n\nRun: https://github.com/lancedb/lance-bench/actions/runs/27729148978\nJob: bench-io / benchmark\n\nAffected benchmarks from the regression analysis:\n\n- \n - Concern: HIGH, p-value 0.001752\n - Older mean: 71.918 ms; last-4 mean: 97.572 ms; change: +35.67% slower\n- \n - Concern: HIGH, p-value 0.009938\n - Older mean: 71.957 ms; last-4 mean: 92.937 ms; change: +29.16% slower\n- \n - Concern: HIGH, p-value 0.007343\n - Older mean: 194.898 ms; last-4 mean: 243.843 ms; change: +25.11% slower\n\nRecent timing history shows the step-up begins with the 2026-06-17 00:13 benchmark result:\n\n\\n\nSuspected commit window: after \ and at/before \ on 2026-06-16/17. Commits in that window include:\n\n- \ fix: merge_insert silently drops matches when a leading payload column is all-null (#7251)\n- \ chore: release beta version 8.0.0-beta.15\n- \ feat: expose session cache key inventory (#7298)\n- \ chore: release beta version 8.0.0-beta.16\n\nDuplicate checks performed before filing:\n\n- Open issues searched by \ and : no matching open issue found.\n- Closed issues searched by \ and : no matching recently closed issue found.\n- Broad open search for \ found #7276, but it covers , not this IO random-read family.\n\nOther regressions in the same run were not filed separately because duplicate/recently closed issues already exist for TPCH scan, IVF-PQ, and decode/encoding families.
The lance-bench run for Lance SHA 6484a0b flagged high/medium-confidence regressions in the Rust IO \ benchmarks.\n\nRun: https://github.com/lancedb/lance-bench/actions/runs/27729148978\nJob: bench-io / benchmark\n\nAffected benchmarks from the regression analysis:\n\n- \n - Concern: HIGH, p-value 0.001752\n - Older mean: 71.918 ms; last-4 mean: 97.572 ms; change: +35.67% slower\n- \n - Concern: HIGH, p-value 0.009938\n - Older mean: 71.957 ms; last-4 mean: 92.937 ms; change: +29.16% slower\n- \n - Concern: HIGH, p-value 0.007343\n - Older mean: 194.898 ms; last-4 mean: 243.843 ms; change: +25.11% slower\n\nRecent timing history shows the step-up begins with the 2026-06-17 00:13 benchmark result:\n\n\\n\nSuspected commit window: after \ and at/before \ on 2026-06-16/17. Commits in that window include:\n\n- \ fix: merge_insert silently drops matches when a leading payload column is all-null (#7251)\n- \ chore: release beta version 8.0.0-beta.15\n- \ feat: expose session cache key inventory (#7298)\n- \ chore: release beta version 8.0.0-beta.16\n\nDuplicate checks performed before filing:\n\n- Open issues searched by \ and : no matching open issue found.\n- Closed issues searched by \ and : no matching recently closed issue found.\n- Broad open search for \ found #7276, but it covers , not this IO random-read family.\n\nOther regressions in the same run were not filed separately because duplicate/recently closed issues already exist for TPCH scan, IVF-PQ, and decode/encoding families.