Right now all JSON produced by these UDFs has the fields in Uppercase, which is weird when using JSON (specially in Spark, using the struct fields). Should we add some struct tags so it's in snake_lower_case?
As I answered into the PR, I'm concerned about backward compatibility. All queries that are using these functions will stop working with no error message.
Right now all JSON produced by these UDFs has the fields in Uppercase, which is weird when using JSON (specially in Spark, using the struct fields). Should we add some struct tags so it's in
snake_lower_case
?WDYT @src-d/data-processing?