Drop de las filas duplicadas pandas excepto nan
df[(~df.duplicated()) | (df['col'].isnull())]
Victorious Vulture
df[(~df.duplicated()) | (df['col'].isnull())]