Schema hints: define column type as struct and incrementally add fields with schema evolution

yit
Databricks Partner

Hey everyone,

I want to set column type as empty struct via schema hints without specifying subfields. Then I expect the struct to be evolved with subfields through schema evolution when new subfields appear in the data. 

But, I've found in the documentation this explanation: 

yit_2-1758029850608.png

Does this affect JSON files as well? Or, can I define empty struct and then evolve it with subfields? 

If yes, how? Because I've tried different approaches but nothing works.