当unicodecsv.DictReader在Python2.7中解析UTF-8-BOM文件时，如何从第一个字段名中删除引号字符？

Question

The issue is when the class unicodecsv.DictReader parses a CSV file's fields when the fields contain quotes and the file is encoded in UTF-8-BOM, the first field retains the quote characters where all consecutive fields have them properly removed.

Example UTF-8-BOM encoded CSV File:

"Field1","Field2","Field3"
content1,content2,content3

Example Python Code:

from unicodecsv import DictReader
filename = "/tmp/test.csv"
with open(filename, mode='r') as read_stream:
     reader = DictReader(read_stream, encoding='utf-8-sig')
     print reader.fieldnames

Print Value:

['"Field1"','Field2','Field3']

Is there a way to have that first field be like the others and have the quote characters removed?

Answer 1

One way is to consume the BOM manually yourself (though I expect the code as written demonstrates an actual bug in the underlying library and should be added to their issues on github ). After consuming the BOM, use the utf-8 codec instead.

# My test code to write a file with a BOM
import io
filename = "/tmp/test.csv"
with io.open('test.csv', 'w', encoding='utf-8-sig') as f:
    f.write(u'''\
"Field1","Field2","Field3"
content1,content2,content3
''')

from unicodecsv import DictReader
with open(filename, mode='r') as read_stream:
     # Consume the BOM
     read_stream.read(3)
     reader = DictReader(read_stream, encoding='utf-8')
     print reader.fieldnames