amogh-jahagirdar opened a new pull request, #10991:
URL: https://github.com/apache/iceberg/pull/10991

   Leaving in draft as we discuss as a community, but this change adds support 
for int/long -> string conversion without any additional metadata changes. 
InclusiveMetricsEvaluator is updated to do a check where if a type is string 
and the bounds are equal and the bounds are either 4/8 bytes, that file is 
assumed to have been type promoted and as a result the file must be read. This 
change also updates the vectorized read path so that the projection from int to 
string can happen after type promotion. Note, ParquetRowGroup filter needed to 
be updated to identify a possible type promotion and ensure the row group is 
read before comparing stats since that would fail due to casting issues.
   
   Leaving in draft as we discuss as a community on the approach and the code 
still needs to be cleaned up/more tests/considerations for other file formats 
like ORC/Avro.


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to