Using the Esri Geodatabase (ArcSDE Geodb) Reader and Writer what would be the best settings to get the best performance to truncate and insert annotation feature class records to a writer. These annotation feature classes are not versioned

I've found for absolute best performance with SDE is to write directly to the underlying database (eg SQL) completely bypassing the SDE layer.

This of course comes with limitations and risks and depends on how you environment is setup

This would be going from an Oracle to SQL server database. I wasn't sure if the Oracle SHAPE columns (Shape, Length, etc...) would equal to the SQL Server Shape Columns.