CurtHagenlocher opened a new pull request, #2534:
URL: https://github.com/apache/arrow-adbc/pull/2534

   I noticed that the ValueAt helper had compiled into some very inefficient 
IL, so I decided to improve its performance by replacing with a simpler switch 
statement. I then realized that in AdbcDataReader, we don't need to do the type 
check on every row because it's the same underlying Arrow type for the entire 
column -- so it's faster to get a delegate once and keep using the same 
delegate for the entire data set. Benchmarking came up with these results:
   
   | Method      | Mean    | Error    | StdDev   |
   |------------ |--------:|---------:|---------:|
   | Original    | 6.263 s | 0.0575 s | 0.0538 s |
   | New ValueAt | 5.400 s | 0.0424 s | 0.0397 s |
   | Getter      | 4.650 s | 0.0281 s | 0.0249 s |
   


-- 
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: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to