Commits


Fatemah Panahi authored and GitHub committed 04bb068402d
PARQUET-2209: [parquet-cpp] Optimize skip for the case that number of values to skip equals page size (#14545) In the current code, we will read this page because we are using > and not >= in the branch that decides to skip the rest of the page. Also includes minor refactoring to reuse the function available_values_current_page(), and use ConsumeBufferedValues() accordingly. Benchmark results for when batch size = 100K and number of values per page = 100K. ``` BEFORE ------------------------------------------------------------------------------- Benchmark Time CPU Iterations ------------------------------------------------------------------------------- REQUIRED 96831 ns 96326 ns 1000 OPTIONAL 623897 ns 621734 ns 1000 REPEATED 1006153 ns 997482 ns 1000 AFTER ------------------------------------------------------------------------------- Benchmark Time CPU Iterations ------------------------------------------------------------------------------- REQUIRED 2175 ns 2164 ns 1000 OPTIONAL 2743 ns 2719 ns 1000 REPEATED 2368 ns 2424 ns 1000 ``` Lead-authored-by: Fatemah Panahi <panahi@google.com> Co-authored-by: Antoine Pitrou <antoine@python.org> Signed-off-by: Antoine Pitrou <antoine@python.org>