Anakin100100 opened a new pull request, #49375:
URL: https://github.com/apache/arrow/pull/49375

   ### Rationale for this change
   Fixing #49310 
   
   ### What changes are included in this PR?
   I added a check for the if else compute kernels that verify that the result 
will fit in the offset of the result. The check asserts that no matter which 
values are chosen from either the left or right arrays/scalars there will be no 
overflow in order to use the UnsafeAppend which skips almost all safety checks.
   
   ### Are these changes tested?
   Yes, There are tests for each of the kernels that include an array that the 
values that would overflow the 32 bit index trigger the check. These tests 
don't actually allocate the data to save on time and compute. There are also 
two tests that verify an overflow for a 32 bit offset type and trigger an error 
and the second one that ensures that for a data type with a 64 bit offset there 
is no overflow error. 
   
   ### Are there any user-facing changes?
   Yes, if the result may not fit in a type that relies on 32 bit offsets the 
user will receive a helpful error message instead of a segfault.


-- 
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