兼容python2和python3的python文字

Question

In python 3 I write

s = '\u0300'

and I get a single unicode character, but that doesn't work in python2, I get a string of various characters.

python2 seems to want:

s = u'\u0300'

so that I must check sys.version_info in order to know if I'm running python 2 or 3 and use the former or the latter accordingly.

Is there a way to write the assignment in a way that works in both versions?

Answer 1

Python 3 (starting from 3.3 IIRC) supports u'' string literals. Or you can use

from __future__ import unicode_literals

to get Python3-compatible string literals in Python 2. This statement is no-op in Python 3.

Answer 2

The most compatible way is to use the six module or a similar compatibility shim:

import six

s = six.u('\u0300')

six is available on PyPI, and is extremely broadly used and well-tested.

Another alternative is:

from __future__ import unicode_literals 

s = '\u0300'

But, this can conflict with a number of Python 2 expectations, that string items can contain bytes. With unicode_literals on, string literals will be silently, almost magically upgraded to Unicode strings. Python culture often favors explicit over implicit.