Create benchmark for sorted data #18976

Activity

1. zhuqi-lucas

   added

   enhancementNew feature or requestNew feature or request

   on Nov 28, 2025
2. zhuqi-lucas

   added the

   Feature

   issue type on Nov 28, 2025
3. zhuqi-lucas

   added a parent issue on Nov 28, 2025

   • Fast parquet order inversion #17172
4. 2010YOUY01 commented on Nov 28, 2025

   

   2010YOUY01

   on Nov 28, 2025

   Member

   More actions

   Just to provide a bit of background:

   Suppose we have a Parquet file with an existing sort order of event_time ASC.
   Now we want to benchmark workloads that read the tail of the file using a query such as:

   select *
   from t
   order by event_time desc
   limit 10

   This is used to evaluate the optimizations in the linked issue/PRs.

   👍React with 👍1Reacted by Qi Zhu
5. zhuqi-lucas commented on Nov 28, 2025

   

   zhuqi-lucas

   on Nov 28, 2025

   ContributorAuthor

   More actions

   Thank you @2010YOUY01 for good note!
6. github-merge-queue

   added a commit that references this issue on Dec 8, 2025

   Add sorted data benchmark. (#19042)

   ...

   cde6dfa
7. zhuqi-lucas

   closed this as completedon Dec 15, 2025
8. zhuqi-lucas

   self-assigned this

   on Dec 25, 2025
9. zhuqi-lucas

   added a commit that references this issue on Jan 7, 2026

   Add sorted data benchmark. (apache#19042)

   ...

   f76f58a