TheR1sing3un commented on PR #7742: URL: https://github.com/apache/paimon/pull/7742#issuecomment-4353261828
Good catch — the previous round drifted from Java despite the description claiming alignment. Did a strict audit against `DataTableBatchScan.applyPushDownLimit` this round and pushed three commits: - `3051ff69a` collapses the loop to mirror Java exactly: only count splits where `mergedRowCount` is present, drop the `row_count` fallback, append at the same point Java does. - `68656b74d` adds the `hasNonPartitionFilter()` short-circuit I'd missed (when the predicate touches a non-partition column, per-split counts are pre-filter and would early-return the loop too low). New helper `_has_non_partition_filter()` mirrors `SnapshotReaderImpl.hasNonPartitionFilter` using the existing `_get_all_fields` walker. - `3878ceac4` trims the comments down to a single docstring reference per method. 11/11 unit tests pass. Apologies for the back-and-forth. -- 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]
