I used this :
u = unicode(text, 'utf-8')
But getting error with Python 3 (or... maybe I just forgot to include something) :
NameError: global name 'unicode' is not defined
Thank you.
5 Answers
Literal strings are unicode by default in Python3.
Assuming that text is a bytes object, just use text.decode('utf-8')
unicode of Python2 is equivalent to str in Python3, so you can also write:
str(text, 'utf-8')
if you prefer.
4 Comments
Gank
TypeError: decoding str is not supported
John La Rooy
@Gank, In Python3 a
str is unicode, ie. it is "decoded" so it makes no sense to call decode on it
Simon
Same TypeError. Please just replace with str(txt), or the code from @magicrebirth below
killua8p
The original sample is not clear. So in python3, if you want to do
str(text, 'utf-8'), text must be a string binary. e.g. str(b'this is a binary', 'utf-8')What's new in Python 3.0 says:
All text is Unicode; however encoded Unicode is represented as binary data
If you want to ensure you are outputting utf-8, here's an example from this page on unicode in 3.0:
b'\x80abc'.decode("utf-8", "strict")
1 Comment
http8086
this is exactly what we need for '\x80abc'.decode("utf-8", "strict") in Python 2, thanks
As a workaround, I've been using this:
# Fix Python 2.x.
try:
UNICODE_EXISTS = bool(type(unicode))
except NameError:
unicode = lambda s: str(s)
3 Comments
nicbou
Why are you using a lambda function? These methods are called the same way in any case. This is a simpler variation:
try: unicode = str; except: pass.
Nickolai
It seems like you can just do
unicode = str since it won't fail in either 2 or 3Nickolai
Or
from six import u as unicode which I'd prefer simply because it's more self-documenting (since six is a 2/3 compatibility layer) than unicode = strThis how I solved my problem to convert chars like \uFE0F, \u000A, etc. And also emojis that encoded with 16 bytes.
example = 'raw vegan chocolate cocoa pie w chocolate & vanilla cream\\uD83D\\uDE0D\\uD83D\\uDE0D\\u2764\\uFE0F Present Moment Caf\\u00E8 in St.Augustine\\u2764\\uFE0F\\u2764\\uFE0F '
import codecs
new_str = codecs.unicode_escape_decode(example)[0]
print(new_str)
>>> 'raw vegan chocolate cocoa pie w chocolate & vanilla cream\ud83d\ude0d\ud83d\ude0d❤️ Present Moment Cafè in St.Augustine❤️❤️ '
new_new_str = new_str.encode('utf-16', errors='surrogatepass').decode('utf-16')
print(new_new_str)
>>> 'raw vegan chocolate cocoa pie w chocolate & vanilla cream😍😍❤️ Present Moment Cafè in St.Augustine❤️❤️ '
1 Comment
Oskar Austegard
For more on 'surrogatepass' see docs.python.org/3/library/codecs.html#error-handlers
In a Python 2 program that I used for many years there was this line:
ocd[i].namn=unicode(a[:b], 'utf-8')
This did not work in Python 3.
However, the program turned out to work with:
ocd[i].namn=a[:b]
I don't remember why I put unicode there in the first place, but I think it was because the name can contains Swedish letters åäöÅÄÖ. But even they work without "unicode".
text.encode('unicode_escape')would be enough I guess