yihua opened a new issue, #917: URL: https://github.com/apache/hudi-rs/issues/917
The record index point lookup from #869 left three small cleanups in `crates/core`: - The scanner test `test_scan_hfile_records_key_predicate_keeps_delete_blocks_whole` says delete blocks must be read whole and asserts that a tombstone for an unrequested key survives. Narrowing a delete block to the requested keys would be equally sound, since the merge is per key, so the test pins an optimization that has not been made rather than the contract. It should assert only that tombstones for requested keys survive. - `lookup_record_index` recomputes the keyed whole-read bound (`hfile_whole_read_max_size(..).min(HFILE_WHOLE_READ_WITH_KEYS_MAX_SIZE)`) that `HFileBaseFileReader::open` also computes. A helper beside the constant in `storage::reader` would keep the two from drifting. - `index::record` imports the metadata table shard layout (file id prefix and format, key hashing, shard count, slice listing) from `write::metadata`, so the read path depends on the write module. Those helpers belong under `metadata`. -- 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]
