Kudu allows case sensitive names, e.g. "t1" and "T1" are separate table names. Impala should eventually follow the standard casing behavior (at least matching postgres), i.e. all identifiers get undercased unless they are quoted.
For Kudu/Impala interactions, Kudu will support case sensitive naming and the behavior in Impala would be:
- non-quoted identifiers are always lower-cased within Impala before calling any Kudu APIs
- quoted identifiers are left as is
- same goes for column names