arrow-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andy Grove (JIRA)" <j...@apache.org>
Subject [jira] [Created] (ARROW-4748) [Rust] [DataFusion] GROUP BY performance could be optimized
Date Sun, 03 Mar 2019 15:25:00 GMT
Andy Grove created ARROW-4748:
---------------------------------

             Summary: [Rust] [DataFusion] GROUP BY performance could be optimized
                 Key: ARROW-4748
                 URL: https://issues.apache.org/jira/browse/ARROW-4748
             Project: Apache Arrow
          Issue Type: Improvement
          Components: Rust, Rust - DataFusion
    Affects Versions: 0.12.0
            Reporter: Andy Grove
             Fix For: 0.13.0


The logic to build the group by keys is row-based, performing an array downcast on every single
group by value. This could be done in a columnar way instead.

 

I also wonder if it is possible to avoid converting the result map to an array of map entries.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message