2010YOUY01 commented on PR #8114:
URL: 
https://github.com/apache/arrow-datafusion/pull/8114#issuecomment-1805340496

   > I wonder if this modification will lead to performance degradation.
   > 
   > Each time we evaluate the function expression, we have to pick an 
implementation based on the input type. Previously, the picking was executed 
only once during compilation/planning.
   
   That's a good point, but I think for this case, the extra overhead is one or 
two function calls per batch(typically 8k rows)
   I tried those functions on TPCH data, and there is no noticeable difference
   ```sql
   select to_timestamp(l_shipdate), to_timestamp_seconds(l_commitdate), 
to_timestamp_millis(l_receiptdate), to_timestamp_micros(l_shipdate), 
to_timestamp_nanos(l_shipdate), from_unixtime(l_orderkey), abs(l_orderkey) from 
lineitem;
   ```


-- 
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