u'somestring'和unicode（'somestring'），python 2.7有什么区别

Question

I was getting 'UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 14: ordinal not in range(128)' when I was concatenating python strings with Django model.CharField like this:

some_variable = unicode("jotain älähti") + self.some_charfield

After I switched to this:

some_variable = u"jotain älähti" + self.some_charfield

It didn't raise the error anymore. What is the difference between u and the unicode function in python? I'm using python 2.7.5 and Django 1.7.1 Why does it not raise the error anymore?

I'm not sure why it would have to decode in the first place. Isn't decoding the process of forming human-readable letters and words from bytes? I would understand decoding in this case if I needed to print it, but I never printed it. Could the decoding relate to somehow to the concatenation process? That in order for the program to concatenate, it needs to decode those both strings, and only after that it can make the concatenation, and then encode those to bytes? I had the coding method input like this in the beginning of the file: # - - coding: utf-8 - -

Answer 1

u"ä" is a Unicode text -- a sequence of Unicode codepoints. It may correspond to different byte sequences depending on a character encoding:

>>> u"ä".encode('utf-8')
'\xc3\xa4'
>>> u"ä".encode('cp1252')
'\xe4'
>>> u"ä".encode('utf-16le')
'\xe4\x00'

The encoding declaration # -*- coding: utf-8 -*- specifies your source code encoding . It just makes sure that b"ä" bytestring literal is interpreted as b'\\xc3\\xa4' byte sequence.

The encoding of your source code has nothing to do with encodings that are used at runtime.

unicode(bytestring) is equivalent to bytestring.decode('ascii') here.

>>> b'\xc3\xa4'.decode('ascii')
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
UnicodeDecodeError: 'ascii' codec can't decode byte 0xc3 in position 0: ordinal not in range(128)

Non-ascii literals characters (such as b"ä" ) in a bytestring are deprecated in Python 3 and unicode type is called str there. You could add from __future__ import unicode_literals at the top to interpret "ä" as a Unicode text on both Python 2 and 3.