On uniqueness in the inverse conductivity problem with local data