Tim-Brooks opened a new pull request, #16227: URL: https://github.com/apache/lucene/pull/16227
When a BinaryColumn reports Density.DENSE, IndexingChain now bulk-consumes values from a BytesRefValuesCursor instead of iterating per-doc tuples, mirroring the dense LongColumn/DictionaryColumn paths. The points pass goes through PointValuesWriter.addDenseNDValues, which stages fixed-width N-D records into the scratch buffer in chunks and drains each with a single writeBytes, rather than one writeBytes per value. Contiguous columns can override BytesRefValuesCursor. fillPackedPoints to fill a chunk with a single arraycopy. Replace the per-writer 1 KB densePointsBuffer with a chain-wide SharedIndexingScratch (4 KB BYTES_SCRATCH_SIZE). RAM is tracked via the chain Counter. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
