Some mysterious problem with the "euro". Let's take a real "French" char. >>> sys.getsizeof('abc' + 'œ') 46 >>> sys.getsizeof(('abc' + 'œ').encode('utf-32')) 37 or a "German" char, ẞẞẞẞẞ >>> sys.getsizeof('abc' + '\N{LATIN CAPITAL LETTER SHARP S}') 46 >>> sys.getsizeof(('abc' + '\N{LATIN CAPITAL LETTER SHARP S}').encode('utf-32')) 37