jvanstraten commented on PR #14434:
URL: https://github.com/apache/arrow/pull/14434#issuecomment-1282621073

   > I'm not sure what rounding behavior we have implemented in Acero and 
therefore the safest thing to do would be to reject all plans that are asking 
for a specific rounding behavior.
   
   Assuming that the arithmetic operations are all ultimately implemented using 
standard library operations, you can query the current mode with 
[`fegetround()`](https://en.cppreference.com/w/cpp/numeric/fenv/feround) (it's 
a global processor state thing). But unless an Arrow user goes out of their way 
to change the mode and this somehow propagates to all threads correctly, it's 
just going to be `FE_TONEAREST`, which corresponds with Substrait `TIE_TO_EVEN` 
(`TIE_AWAY_FROM_ZERO` does not seem to be implemented by the standard library, 
the rest of the mapping should be fairly straightforward).


-- 
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-unsubscr...@arrow.apache.org

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

Reply via email to