When dumping large table it would be nice the tool writes the CSV file incremental so lets say every 1000 rows dumped get appended to the file so external tools can already process partial data.