BeanPropertyRowMapper traditionally maps lower-case and underscore-based names for each discovered property.
While it allows for customizing those specific names, it does not allow for totally custom mapped names, e.g. derived from a custom annotation.
See the use case described in the tests for #30180.