[GitHub] [arrow] pitrou commented on pull request #35280: GH-29887: [C++] Implement dictionary array sorting

["pitrou (via GitHub)" <gi...@apache.org>]

pitrou commented on PR #35280:
URL: https://github.com/apache/arrow/pull/35280#issuecomment-1550047309

   Just for the record, it's a pity that the all-or-mostly-null case is a bit slower:
   ```
   ArraySortIndicesInt64WideDict/32768/10000         254 us          254 us         2756 bytes_per_second=122.985M/s items_per_second=16.1199M/s null_percent=0.01 size=32.768k
   ArraySortIndicesInt64WideDict/32768/100           272 us          272 us         2615 bytes_per_second=115.005M/s items_per_second=15.0739M/s null_percent=1 size=32.768k
   ArraySortIndicesInt64WideDict/32768/10            324 us          324 us         2162 bytes_per_second=96.5235M/s items_per_second=12.6515M/s null_percent=10 size=32.768k
   ArraySortIndicesInt64WideDict/32768/2             386 us          386 us         1811 bytes_per_second=80.9335M/s items_per_second=10.6081M/s null_percent=50 size=32.768k
   ArraySortIndicesInt64WideDict/32768/1             357 us          357 us         1956 bytes_per_second=87.483M/s items_per_second=11.4666M/s null_percent=100 size=32.768k
   ArraySortIndicesInt64WideDict/32768/0             254 us          254 us         2751 bytes_per_second=122.957M/s items_per_second=16.1163M/s null_percent=0 size=32.768k
   ArraySortIndicesInt64WideDict/1048576/100        3398 us         3396 us          205 bytes_per_second=294.442M/s items_per_second=38.593M/s null_percent=1 size=1048.58k
   ArraySortIndicesInt64WideDict/8388608/100       31369 us        31341 us           22 bytes_per_second=255.253M/s items_per_second=33.4566M/s null_percent=1 size=8.38861M
   ```
   
   This can be left to another issue or PR, though. Overall the improvement is very nice, especially for strings.


-- 
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: github-unsubscribe@arrow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org