query.py 文件源码

def csv_safe_unicode(row, ignoreErrors=True):
    """
    Given an array of values, make sure all strings are unicode in Python 3 and
    str in Python 2.  The csv.writer in Python 2 does not handle unicode
    strings and in Python 3 it does not handle byte strings.

    :param row: an array which could contain mixed types.
    :param ignoreErrors: if True, convert what is possible and ignore errors.
        If false, conversion errors will raise an exception.
    :returns: either the original array if safe, or an array with all byte
        strings converted to unicode in Python 3 or str in Python 2.
    """
    # This is unicode in Python 2 and bytes in Python 3
    wrong_type = six.text_type if str == six.binary_type else six.binary_type
    # If all of the data is not the wrong type, just return it as is.  This
    # allows non-string types, such as integers and floats.
    if not any(isinstance(value, wrong_type) for value in row):
        return row
    # Convert the wrong type of string to the type needed for this version of
    # Python.  For Python 2 unicode gets encoded to str (bytes).  For Python 3
    # bytes get decoded to str (unicode).
    func = 'encode' if str == six.binary_type else 'decode'
    row = [getattr(value, func)('utf8', 'ignore' if ignoreErrors else 'strict')
           if isinstance(value, wrong_type) else value for value in row]
    return row