Why do some strings encode in utf-16, while others only encode in utf-8?

Question

>>> unicode('восстановление информации', 'utf-16')
Traceback (most recent call last):
File "<interactive input>", line 1, in <module>
File "C:\Python27\lib\encodings\utf_16.py", line 16, in decode
return codecs.utf_16_decode(input, errors, True)
UnicodeDecodeError: 'utf16' codec can't decode byte 0xb8 in position 48: truncated data

>>> unicode('восстановление информации', 'utf-8')
u'\u0432\u043e\u0441\u0441\u0442\u0430\u043d\u043e\u0432\u043b\u0435\u043d\u0438\u0435 \u0438\u043d\u0444\u043e\u0440\u043c\u0430\u0446\u0438\u0438'

Why do these Russian words encode in UTF-8 fine, but not UTF-16?

Answer 1

You are asking the unicode function to decode a byte string and then giving it the wrong encoding.

Pasting your string into Python-2.7 on OS-X gives

>>> 'восстановление информации'
'\xd0\xb2\xd0\xbe\xd1\x81\xd1\x81\xd1\x82\xd0\xb0\xd0\xbd\xd0\xbe\xd0\xb2\xd0\xbb\xd0\xb5\xd0\xbd\xd0\xb8\xd0\xb5 \xd0\xb8\xd0\xbd\xd1\x84\xd0\xbe\xd1\x80\xd0\xbc\xd0\xb0\xd1\x86\xd0\xb8\xd0\xb8'

At this stage it is already a UTF-8 encoded string (probably your terminal determined this), so you can decode it by specifying the utf-8 codec

>>> 'восстановление информации'.decode('utf-8')
u'\u0432\u043e\u0441\u0441\u0442\u0430\u043d\u043e\u0432\u043b\u0435\u043d\u0438\u0435 \u0438\u043d\u0444\u043e\u0440\u043c\u0430\u0446\u0438\u0438'

But not UTF-16 as that would be invalid

>>> 'восстановление информации'.decode('utf-16')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/encodings/utf_16.py", line 16, in decode
    return codecs.utf_16_decode(input, errors, True)
UnicodeDecodeError: 'utf16' codec can't decode byte 0xb8 in position 48: truncated data

If you want to encode a unicode string to UTF-8 or UTF-16, then use

>>> u'восстановление информации'.encode('utf-16')
'\xff\xfe2\x04>\x04A\x04A\x04B\x040\x04=\x04>\x042\x04;\x045\x04=\x048\x045\x04 \x008\x04=\x04D\x04>\x04@\x04<\x040\x04F\x048\x048\x04'
>>> u'восстановление информации'.encode('utf-8')
'\xd0\xb2\xd0\xbe\xd1\x81\xd1\x81\xd1\x82\xd0\xb0\xd0\xbd\xd0\xbe\xd0\xb2\xd0\xbb\xd0\xb5\xd0\xbd\xd0\xb8\xd0\xb5 \xd0\xb8\xd0\xbd\xd1\x84\xd0\xbe\xd1\x80\xd0\xbc\xd0\xb0\xd1\x86\xd0\xb8\xd0\xb8'

Notice the input strings are unicode (have a u at the front), but the outputs here are byte-strings (they don't have u at the start) which contain the unicode data encoded in the respective formats.