UnicodeEncodeError: 'charmap' codec can't encode characters in position 0-14: character maps to <undefined>

Hi! How to solve "UnicodeEncodeError: ‘charmap’ codec can’t encode characters in position 0-14: character maps to "

for this code:
stockList = ts.get_stock_basics()

Hi Jonathan,

Sorry you have not provided enough information to solve the problem. I’m
going to have to do a lot of guessing.

The absolute minimum we would need to see would be the entire
traceback, showing the full error starting with the line “Traceback” and
ending with the error message. But even that may not be enough.

What is “ts”? Where did it come from? What does it do?

Why is it using the ‘charmap’ codec? Can you tell it to use a different
encoding instead of ‘charmap’?

Are you reading data from a website? What encoding is the website using?
If you’re unsure, it is probably using UTF-8. Can you tell the mystery
ts object to use encoding=‘utf-8’?

You could try running your script on an Apple Mac or Linux instead of
Windows, that could fix the problem. (The ‘charmap’ codec is only used
on Windows.)

I tried utf-8 in the code like this before: stockList = ts.get_stock_basics(encoding= “utf-8”) . but gave me other error message. I don’t know how to put the encoding.

import pymysql
import sys
import tushare as ts

db = pymysql.connect(host=“localhost”,user=“root”,password=“cixlerui83466”,database=‘pythonstock’,port=3306)
print(‘Error whaen Connecting to DB.’)
cursor = db.cursor()
stockList = ts.get_stock_basics()

#for code in stockList.index:

    #createSql='CREATE TABLE stock_'+code+'( date varchar(255) , open float, close float, high float, low float, vol int(11))'
    #print('Error when Creating table for:' + code)


Traceback (most recent call last):
File “D:\Users\jon_j\eclipse-workspace\HelloP8\src\CreateTable.py”, line 11, in
stockList = ts.get_stock_basics()
File “D:\Python310\lib\site-packages\tushare\stock\fundamental.py”, line 52, in get_stock_basics
File “D:\Python310\lib\encodings\cp1252.py”, line 19, in encode
return codecs.charmap_encode(input,self.errors,encoding_table)[0]
UnicodeEncodeError: ‘charmap’ codec can’t encode characters in position 0-14: character maps to

In Python 3.6+, output to a Windows console file uses Unicode. All other files, however, default to using the process ANSI code page, which is typically a legacy encoding. The latter includes the standard files (i.e. stdin, stdout, stderr) when they’re redirected to something other than a console, such as a disk file or pipe. In your case the encoding of sys.stdout is code page 1252, maybe because the output is piped to Eclipse. There are two environment variables that you can set to override the encoding of sys.stdout. To force just the standard files to use UTF-8 , set PYTHONIOENCODING=utf-8. To use UTF-8 as the default in all cases, enable UTF-8 mode via PYTHONUTF8=1. To configure the latter for a single running instance of Python, use the command-line option -X utf8 instead of setting PYTHONUTF8=1.

Jonathan said:

“I tried utf-8 in the code like this before: stockList =
ts.get_stock_basics(encoding= “utf-8”) . but gave me other error

Don’t be shy. Copy and paste the error message so we can see it.

I don’t know how to put the encoding.

What does the tushare documentation say? I can only find these, which
are all in Chinese:

You could try Google Translate.

Looking at the source code here:

the get_stock_basics() function is hard-coded to use the GBK encoding:

text = text.decode('GBK')

so at this point I would say it is a bug in the tushare library.

The get_stock_basics() call prints “本接口即将停止更新,请尽快使用Pro版接口:https​://waditu.com/document/2”, which translates as “This interface will stop updating, please use the Pro version interface as soon as possible: https​://waditu.com/document/2”. Here’s the source from PyPI:

def get_stock_basics(date=None):

    wdate = du.last_tddate() if date is None else date
    wdate = wdate.replace('-', '')
    if wdate < '20160809':
        return None
    datepre = '' if date is None else wdate[0:4] + wdate[4:6] + '/'
    request = Request(ct.ALL_STOCK_BASICS_FILE%(datepre, '' if date is None else wdate))
    text = urlopen(request, timeout=10).read()
    text = text.decode('GBK')
    text = text.replace('--', '')
    df = pd.read_csv(StringIO(text), dtype={'code':'object'})
    df = df.set_index('code')
    return df

The hard-coded GBK decode() in the source is unrelated to the encoding error. One can reasonably assume that this is a known encoding for the text on the website that’s being accessed via urlopen().

The problem is just the print() call and the encoding of sys.stdout. The Chinese text that it tries to print can’t be encoded with code page 1252, which is an 8-bit encoding for Western European alphabets. Refer to my previous post for a few simple ways to force sys.stdout to use UTF-8.

Nicely spotted. Thanks Eryk Sun.