Commits


mwish authored and GitHub committed 14a433aa65d
GH-37873: [C++][Parquet] DELTA_BYTE_ARRAY: avoid copying data when possible (#37874) ### Rationale for this change When decoding DELTA_BYTE_ARRAY data, if the prefix (respectively suffix) is empty, we don't need to recreate the original string by copying the data into a new buffer, we can just point to the existing suffix (respectively suffix). ### What changes are included in this PR? Avoid spurious memory copies in the DELTA_BYTE_ARRAY decoder (also reducing the memory footprint when decoding). Benchmark numbers show that decoding can be up to 2x faster. ### Are these changes tested? Yes, already tested. ### Are there any user-facing changes? No. * Closes: #37873 Lead-authored-by: mwish <maplewish117@gmail.com> Co-authored-by: Antoine Pitrou <antoine@python.org> Signed-off-by: Antoine Pitrou <antoine@python.org>