[Design Discussion] Case Sensitivity Handling in Thunder

[hwupathum]

Thunder supports multiple relational databases as storage backends. Different databases have different default behaviors for string comparison—some are case sensitive, others are case insensitive, and some allow configuration through collation settings.

As a result, queries that filter or match values may produce different results depending on which database is used. The same query could behave differently across deployments, leading to inconsistent behavior at the storage layer.

This creates an architectural concern around how Thunder should handle case sensitivity when interacting with multiple databases, and whether there should be any expectation of uniform behavior across database backends.