Hi Werners, the issue is not Databricks (all of the data looks fine and in proper encoding when I look at it there), the issue is the SIMBA JDBC Driver which by default appears to be bringing over columns of datatype STRING as SQL_VARCHAR instead of SQL_WVARCHAR, for this specific use case i need to use the table interface. Other JDBC drivers typically have some property you can set to tell it to use unicode and UTF8, shocked I can't find this for the SIMBA JDBC Driver which Databricks provides on the site.