JingsongLi commented on PR #1164: URL: https://github.com/apache/paimon-rust/pull/1164#issuecomment-6606998202
Requirement fit: SUPPORTED. Implementation: CLEAN at 7270d1e1. The earlier repeated-NULL-key finding is addressed by null-safe comparisons both within and across batches; no new implementation finding in this pass. Validation in an isolated worktree with Rust 1.98.0: - 50 KV reader unit tests and 37 table-level single-run / nullable-key / first-row / user-sequence / audit-log tests passed. - An independent persisted-Parquet probe reconstructs a historical file containing repeated NULL keys with accurate file size, row counts and statistics, commits it plus a disjoint second file, and reads through the ordinary table planner. It returns exactly one NULL-key row and the disjoint row for batch sizes 1 and 64 under deduplicate and first-row (using first-row's public all-files scan mode for level-0 files). The planner is asserted to produce a two-file non-raw split. - Reviewed comparator parity, retract handling, fallback schema/level columns, lookup and audit isolation. Local merge-tree integration with current main 062f1bcd is conflict-free. Production readiness remains blocked by the exact head's [CI check job](https://github.com/apache/paimon-rust/actions/runs/41451513236/job/125297582976): dependency policy rejects inherited thrift 0.17.0 for RUSTSEC-2026-0435. This PR does not change that dependency. Current main has upgraded the Arrow/Parquet stack and no longer contains thrift in Cargo.lock; please rebase/merge that baseline and obtain green required CI before merging for production. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
