fallintoplace opened a new pull request, #2225: URL: https://github.com/apache/iceberg-go/pull/2225
## What - Reuse batches when requested row-lineage columns are already complete. - Advance the row-position cursor across the batch in one step. ## Why - Avoid rebuilding valid `_row_id` and `_last_updated_sequence_number` Arrow columns. ## Implementation - Retain the input batch when no values need synthesis. - Keep synthesis for null values and rebuild only columns that need it. ## Benchmark - Go 1.26.3, Apple M1 Pro, 10 runs; row-lineage batch helper. - 1,024 rows: complete 15.7 us to 195 ns; row_id_sparse 15.7 us to 9.58 us; both_sparse 15.6 us to 16.8 us. - 65,536 rows: complete 788 us to 190 ns; row_id_sparse 782 us to 467 us; both_sparse 781 us to 785 us. - Complete 65,536-row batches: 2.15 MiB to 352 B and 29 to 2 allocations. -- 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]
