问题
StopWords = pd.read_csv('stopwords.csv',encoding='UTF-8', quotechar='|',names=['StopWords'])
I am trying to read a CSV file that contains Persian language text, and this is the error I get:
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte
回答1:
Without seeing the binary content of the file it is difficult to guess the actual encoding but UTF-8, with or without a BOM (Byte order Marker) cannot start with an 0xFF.
If it starts with an 0xFF, then that would suggest that it is probably in Little Endian UTF-16 to UTF-32 which are the only Unicode serialisations that have a byte order marker starting with 0xFF.
https://en.wikipedia.org/wiki/Byte_order_mark gives some explanation.
It is also possible that it is a Persian specific character set. National character sets should be avoided if a Unicode option is available, for the generation of your source CSV files.
来源：https://stackoverflow.com/questions/58199571/unicodedecodeerror-utf-8-codec-cant-decode-byte-0xff-in-position-0-invalid