andygrove commented on PR #2659: URL: https://github.com/apache/datafusion-comet/pull/2659#issuecomment-3462492793
> > Comet always unpacks dictionaries in ScanExec > > Just for me to understand this, is it true for all scan types such as `native_comet`, `native_datafusion`, `native_iceberg_compat`, and maybe `iceberg-rust`? It is true for the two scans we support today (`native_comet` and `native_iceberg_compat`), because both are hybrid scans where the native plan reads batches from the JVM via `ScanExec`. If/when we support the new experimental scans, we will need to ensure they unpack dictionary-encoded arrays if they do not already. I am guessing they already do this, since they are based on DataFusion `DataSourceExec`, where `UTF8View` is now preferred over dictionary-encoding, but @mbutrovich can confirm. -- 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]
