Since CSV is based on arbitrary columns rather than keys and values, we always hesitate to alter the data within a CSV file because every change requires anyone doing automated imports of the data into other programs to update their setup. We typically only update the CSV output when we release a major update to the API. The last update was several years ago.