62

I have a big amount of files and parser. What I Have to do is strip all non utf-8 symbols and put data in mongodb. Currently I have code like this. 

with open(fname, "r") as fp:
    for line in fp:
        line = line.strip()
        line = line.decode('utf-8', 'ignore')
        line = line.encode('utf-8', 'ignore')

somehow I still get an error

bson.errors.InvalidStringData: strings in documents must be valid UTF-8: 
1/b62010montecassianomcir\xe2\x86\x90ta0\xe2\x86\x90008923304320733/290066010401040101506055soccorin

I don't get it. Is there some simple way to do it?

UPD: seems like Python and Mongo don't agree about definition of Utf-8 Valid string.

maudulus
  • 10,627
  • 10
  • 78
  • 117
Darth Kotik
  • 2,261
  • 1
  • 20
  • 29

5 Answers5

94

Try below code line instead of last two lines. Hope it helps:

line=line.decode('utf-8','ignore').encode("utf-8")
Irshad Bhat
  • 8,479
  • 1
  • 26
  • 36
  • 1
    I have some invisible chars which are appearing only after insertion. any fix for this? – user168983 Sep 29 '16 at 08:04
  • 1
    @user168983 can you give an example – Irshad Bhat Sep 29 '16 at 19:41
  • 5
    this `line.decode('utf-8','ignore').encode("utf-8")` produce this error _AttributeError: 'str' object has no attribute 'decode'_, i work with python3 – Chedi Bechikh Jan 20 '17 at 22:22
  • 24
    @ChediBechikh Here is how you do it in python3 `bytes(line, 'utf-8').decode('utf-8','ignore')` – Irshad Bhat Jan 20 '17 at 22:46
  • this does not change anything, the error exist when printing so i write this code `print(ligne.rstrip().encode('utf-8','ignore')) ` but the problem with encode is , it add a b caracter like **b'press'** – Chedi Bechikh Jan 20 '17 at 23:19
  • 2
    This doesn't seem to work. I get lots of special chars: `\00\00\00\00\00` – dengar81 Feb 01 '21 at 11:27
  • It looks like the only difference between the code in the question and the answer is that encode() no longer has ignore. Why would this make a difference? The default is strict which would raise an exception rather than skip the invalid utf8? – Richard EB Dec 07 '21 at 20:42
37

For python 3, as mentioned in a comment in this thread, you can do:

line = bytes(line, 'utf-8').decode('utf-8', 'ignore')

The 'ignore' parameter prevents an error from being raised if any characters are unable to be decoded.

If your line is already a bytes object (e.g. b'my string') then you just need to decode it with decode('utf-8', 'ignore').

Alex
  • 12,078
  • 6
  • 64
  • 74
7

Example to handle no utf-8 characters

import string

test=u"\n\n\n\n\n\n\n\n\n\n\n\n\n\nHi <<First Name>>\nthis is filler text \xa325 more filler.\nadditilnal filler.\n\nyet more\xa0still more\xa0filler.\n\n\xa0\n\n\n\n\nmore\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nfiller.\x03\n\t\t\t\t\t\t    almost there \n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nthe end\n\n\n\n\n\n\n\n\n\n\n\n\n"

print ''.join(x for x in test if x in string.printable)
Hafiz Muhammad Shafiq
  • 8,168
  • 12
  • 63
  • 121
  • 11
    this removes all non-ascii characters, which includes many, many valid UTF-8 characters – szxk Jun 08 '17 at 18:08
0
with open(fname, "r") as fp:
for line in fp:
    line = line.strip()
    line = line.decode('cp1252').encode('utf-8')
Willem
  • 1,304
  • 1
  • 8
  • 7
  • 3
    This will be horribly wrong if the original input encoding is not in fact code page 1252. You remove the error, but produce garbage. The error is there for a reason, to prevent you from producing garbage. – tripleee Nov 09 '20 at 09:10
0

I would just encode in ascii.

line = line.encode('ascii', 'ignore')
Bojan S.
  • 175
  • 1
  • 5