UnicodeEncodeError when redirecting stdout

Question

I'm having a problem regarding Unicode in Python. I can print the output fine in a regular terminal, but if I redirect stdout elsewhere (or capture it with the subprocess module), I get a UnicodeEncodeError:

$ cat example.py 
print u'Example: \u00F1'
$ python example.py 
Example: ñ
$ python example.py > /dev/null
Traceback (most recent call last):
  File "example.py", line 1, in <module>
    print u'Example: \u00F1'
UnicodeEncodeError: 'ascii' codec can't encode character u'\xf1' in position 9: ordinal not in range(128)

Why is this? How can I fix it?

http://stackoverflow.com/questions/492483/setting-the-correct-encoding-when-piping-stdout-in-python is essentially the same question, the answers solved the problem I was having (hopefully).. — dbr, Feb 09 '10 at 01:17

score 9 · Accepted Answer · answered Feb 08 '10 at 19:10

9

Pipes that don't lead to the terminal don't have an encoding, therefore you'll need to check sys.stdout.isatty() and encode if needed.

answered Feb 08 '10 at 19:10

Ignacio Vazquez-Abrams

776,304
153
1,341
1,358

4

Or (even better) just always encode. – Thomas Wouters Feb 08 '10 at 19:11
More info on how to define encoding: http://www.python.org/dev/peps/pep-0263/ – Jason Coon Feb 08 '10 at 19:26
1

No, PEP 263 is about specifying *source* encoding, and has nothing to do with output encodings. – Thomas Wouters Feb 08 '10 at 20:39
3

Is there an easy way to "just always encode"? – eric.frederich Feb 08 '11 at 16:05
3

@eric, `print mystring.encode(encoding, 'ignore')` – Cerin Dec 21 '11 at 17:25
4

@eric.frederich Alternatively use this at the start of your script: `sys.stdout = codecs.getwriter("UTF-8")(sys.stdout)`, after importing `codecs` and `sys` of course. – markshep Jul 04 '12 at 00:45

UnicodeEncodeError when redirecting stdout

1 Answers1

Linked