CODEWITHSUNDEEP
pythonunicodedecode
Rob Lund
Rob Lund

Reputation: 603

Unable to decode unicode string in Python 2.4

This is in python 2.4. Here is my situation. I pull a string from a database, and it contains an umlauted 'o' (\xf6). At this point if I run type(value) it returns str. I then attempt to run .decode('utf-8'), and I get an error ('utf8' codec can't decode bytes in position 1-4).

Really my goal here is just to successfully make type(value) return unicode. I found an earlier question that had some useful information, but the example from the picked answer doesn't seem to run for me. Is there something I am doing wrong here?

Here is some code to reproduce:

Name = 'w\xc3\xb6rner'.decode('utf-8')
file.write('Name: %s - %s\n' %(Name, type(Name)))

I never actually get to the write statement, because it fails on the first statement.

Thank you for your help.

Edit:

I verified that the DB's charset is utf8. So in my code to reproduce I changed '\xf6' to '\xc3\xb6', and the failure still occurs. Is there a difference between 'utf-8' and 'utf8'?

The tip on using codecs to write to a file is handy (I'll definitely use it), but in this scenario I am only writing to a log file for debugging purposes.

Upvotes: 5

Views: 17550

Answers (4)

bobince
bobince

Reputation: 536597

So in my code to reproduce I changed '\xf6' to '\xc3\xb6', and the failure still occurs

Not in the first line it doesn't:

>>> 'w\xc3\xb6rner'.decode('utf-8')
u'w\xf6rner'

The second line will error out though:

>>> file.write('Name: %s - %s\n' %(Name, type(Name)))
UnicodeEncodeError: 'ascii' codec can't encode character u'\xf6' in position 7: ordinal not in range(128)

Which is entirely what you'd expect, trying to write non-ASCII Unicode characters to a byte stream. If you use Jiri's suggestion of a codecs-wrapped stream you can write Unicode directly, otherwise you will have to re-encode the Unicode string into bytes manually.

Better, for logging purposes, would be simply to spit out a repr() of the variable. Then you don't have to worry about Unicode characters being in there, or newlines or other unwanted characters:

name= 'w\xc3\xb6rner'.decode('utf-8')
file.write('Name: %r\n' % name)

Name: u'w\xf6rner'

Upvotes: 2

vartec
vartec

Reputation: 134641

It's obviously 1-byte encoding. 'ö' in UTF-8 is '\xc3\xb6'.

The encoding might be:

  • ISO-8859-1
  • ISO-8859-2
  • ISO-8859-13
  • ISO-8859-15
  • Win-1250
  • Win-1252

Upvotes: 5

Jiri
Jiri

Reputation: 16625

Your string is not in UTF8 encoding. If you want to 'decode' string to unicode, your string must be in encoding you specified by parameter. I tried this and it works perfectly:

print 'w\xf6rner'.decode('cp1250')

EDIT

For writing unicode strings to the file you can use codecs module:

import codecs
f = codecs.open("yourfile.txt", "w", "utf8")
f.write( ... )

It is handy to specify encoding of the input/output and using 'unicode' string throughout your code without bothering of different encodings.

Upvotes: 10

Staale
Staale

Reputation: 27998

You need to use "ISO-8859-1":

Name = 'w\xf6rner'.decode('iso-8859-1')
file.write('Name: %s - %s\n' %(Name, type(Name)))

utf-8 uses 2 bytes for escaping anything outside ascii, but here it's just 1 byte, so iso-8859-1 is probably correct.

Upvotes: 3

Related Questions