I'm evaluating if I can use Maria's JDBC connector to join tables from Apache Pinot (they have a prototype jdbc connector). 

When external tables are used (like JDBC or FederatedX), are group bys pushed down or do these connectors return the raw rows to Maria?  I see comments about lack of filter pushdowns for FederatedX.  I wouldn't the raw level data coming back from Pinot (there are too many rows to be performant).

Thanks!

- Dan