Python "SyntaxError: Non-ASCII character '\xe2' in file"

Question

I am writing some python code and I am receiving the error message as in the title, from searching this has to do with the character set.

Here is the line that causes the error

hc = HealthCheck("instance_health", interval=15, target808="HTTP:8080/index.html")

I cannot figure out what character is not in the ANSI ASCII set? Furthermore searching "\xe2" does not give anymore information as to what character that appears as. Which character in that line is causing the issue?

I have also seen a few fixes for this issue but I am not sure which to use. Could someone clarify what the issue is (python doesn't interpret unicode unless told to do so?), and how I would clear it up properly?

EDIT: Here are all the lines near the one that errors

def createLoadBalancer():
    conn = ELBConnection(creds.awsAccessKey, creds.awsSecretKey)
    hc = HealthCheck("instance_health", interval=15, target808="HTTP:8080/index.html")
    lb = conn.create_load_balancer('my_lb', ['us-east-1a', 'us-east-1b'],[(80, 8080, 'http'), (443, 8443, 'tcp')])
    lb.configure_health_check(hc)
    return lb

Answer 1

If you are just trying to use UTF-8 characters or don't care if they are in your code, add this line to the top of your .py file

# -*- coding: utf-8 -*-

Answer 2

You've got a stray byte floating around. You can find it by running

with open("x.py") as fp:
    for i, line in enumerate(fp):
        if "\xe2" in line:
            print i, repr(line)

where you should replace "x.py" by the name of your program. You'll see the line number and the offending line(s). For example, after inserting that byte arbitrarily, I got:

4 "\xe2        lb = conn.create_load_balancer('my_lb', ['us-east-1a', 'us-east-1b'],[(80, 8080, 'http'), (443, 8443, 'tcp')])\n"

Answer 3

Or you could just simply use:

# coding: utf-8

at top of .py file

Answer 4

\xe2 is the '-' character, it appears in some copy and paste it uses a different equal looking '-' that causes encoding errors. Replace the '-'(from copy paste) with the correct '-' (from you keyboard button).

Answer 5

Change the file character encoding,

put below line to top of your code always

# -*- coding: utf-8 -*-

Answer 6

I had the same error while copying and pasting a comment from the web

For me it was a single quote (') in the word

I just erased it and re-typed it.

Answer 7

Adding # coding=utf-8 line in first line of your .py file will fix the problem.

Please read more about the problem and its fix on below link, in this article problem and its solution is beautifully described : https://www.python.org/dev/peps/pep-0263/

Answer 8

I got this error for characters in my comments (from copying/pasting content from the web into my editor for note-taking purposes).

To resolve in Text Wrangler:

1. Highlight the text
2. Go the the Text menu
3. Select "Convert to ASCII"

Answer 9

Based on PEP 0263 -- Defining Python Source Code Encodings

Python will default to ASCII as standard encoding if no other
encoding hints are given.

To define a source code encoding, a magic comment must
be placed into the source files either as first or second
line in the file, such as:

      # coding=<encoding name>

or (using formats recognized by popular editors)

      #!/usr/bin/python
      # -*- coding: <encoding name> -*-

or

      #!/usr/bin/python
      # vim: set fileencoding=<encoding name> :

Answer 10

I had the same issue and just added this to the top of my file (in Python 3 I didn't have the problem but do in Python 2

#!/usr/local/bin/python
# coding: latin-1

Answer 11

If it helps anybody, for me that happened because I was trying to run a Django implementation in python 3.4 with my python 2.7 command

Answer 12

I my case \xe2 was a ’ which should be replaced by '.

In general I recommend to convert UTF-8 to ASCII using e.g. https://onlineasciitools.com/convert-utf8-to-ascii

However if you want to keep UTF-8 you can use

#-*- mode: python -*-
# -*- coding: utf-8 -*-

Answer 13

After about a half hour of looking through stack overflow, It dawned on me that if the use of a single quote " ' " in a comment will through the error:

SyntaxError: Non-ASCII character '\xe2' in file

After looking at the traceback i was able to locate the single quote used in my comment.

Answer 14

I had this exact issue running the simple .py code below:

import sys
print 'version is:', sys.version

DSM's code above provided the following:

1 'print \xe2\x80\x98version is\xe2\x80\x99, sys.version'

So the issue was that my text editor used SMART QUOTES, as John Y suggested. After changing the text editor settings and re-opening/saving the file, it works just fine.

Answer 15

I am trying to parse that weird windows apostraphe and after trying several things here is the code snippet that works.

def convert_freaking_apostrophe(self,string):

   try:
      issuer_rename = string.decode('windows-1252')
   except:
      issuer_rename = string.decode('latin-1')
   issuer_rename = issuer_rename.replace(u'’', u"'")
   issuer_rename = issuer_rename.encode('ascii','ignore')
   try:
      os.rename(directory+"/"+issuer,directory+"/"+issuer_rename)
      print "Successfully renamed "+issuer+" to "+issuer_rename
      return issuer_rename
   except:
      pass

#HANDLING FOR FUNKY APOSTRAPHE
if re.search(r"([\x90-\xff])", issuer):
   issuer = self.convert_freaking_apostrophe(issuer)

Answer 16

I fixed this using pycharm. At the bottom of pycharm you can see file encoding. I noticed that it is UT-8. I changed it to US-ASCII

Answer 17

I had the same issue but it was because I copied and pasted the string as it is. Later when I manually typed the string as it is the error vanished.

I had the error due to the - sign. When I replaced it with manually inputting a - the error was solved.

Copied string 10 + 3 * 5/(16 − 4)

Manually typed string 10 + 3 * 5/(16 - 4)

you can clearly see there is a bit of difference between both the hyphens.

I think it's because of the different formatting used by different OS or maybe just different software.

Answer 18

For me the problem had caused due to "’" that symbol in the quotes. As i had copied the code from a pdf file it caused that error. I just replaced "’" by this "'".

Answer 19

If you want to spot what character caused this just assign the problematic variable to a string and print it in a iPython console.

In my case

In [1]: array = [[24.9, 50.5]​, [11.2, 51.0]]        # Raises an error

In [2]: string = "[[24.9, 50.5]​, [11.2, 51.0]]"     # Manually paste the above array here

In [3]: string
Out [3]: '[[24.9, 50.5]\xe2\x80\x8b, [11.2, 51.0]]' # Here they are!

Answer 20

for me, the problem was caused by typing my code into Mac Notes and then copied it from Mac Notes and pasted into my vim session to create my file. This made my single quotes the curved type. to fix it I opened my file in vim and replaced all my curved single quotes with the straight kind, just by removing and retyping the same character. It was Mac Notes that made the same key stroke produce the curved single quote.

Answer 21

I was unable to find what's the issue for long but later I realised that I had copied a line "UTC-12:00" from web and the hyphen/dash in this was causing the problem. I just wrote this "-" again and the problem got resolved.

So, sometimes the copy pasted lines also give errors. In such cases, just re-write the copy pasted code and it works. On re-writing, it would look like nothing got changed but the error will be gone.

Answer 22

Plenty of good solutions here.

One challenge not really addressed in any of them is how to visually identify certain hard-to-spot non-ASCII characters that resemble other plain ASCII ones. For example, en dashes can appear almost exactly like hyphens and curly quotes look a lot like straight quotes, depending on your text editor's font.

This one-liner, which should work on Mac or Linux, will strip characters not in the ASCII printable range and show you the differences side-by-side:

# assumes Bash shell; for Bourne shell (sh), rearrange as a pipe and
# give '-' as second argument to 'sdiff' instead
sdiff --suppress-common-lines script.py <(tr -cd '\11\12\15\40-\176' <script.py)

The characters \11, \12, and \15 are tab, newline, and carriage return, respectively, in octal; the remaining range is the visible ASCII characters. (hat tip)

Another tip gleaned from this SO thread uses an inverse character class consisting of anything not in the ASCII visible range, and highlights it:

grep --color '[^ -~]' script.py

This should also work fine with the macOS / BSD version of grep.

Answer 23

When I have a similar issue when reading text files i use...

f = open('file','rt', errors='ignore')