Python 3：删除文件中最后一个字符的惯用方法

Question

有没有一种惯用的方法来做到这一点？我刚刚从 Python 2 升级到 Python 3，我正在尝试移植我的脚本，我不得不说我没有留下深刻的印象。据我所知，我的代码可以运行了

从这里

# Not allowed by Python 3 anymore without being in binary mode.
card_names_file.seek(-1, os.SEEK_END)
if card_names_file.read() == ',':
    card_names_file.truncate()

到这里

# Go to end of file just to get position index. lawl.
card_names_file.seek(0, os.SEEK_END)
# Create a temporary just to store said index. More lawl.
eof = card_names_file.tell()
# Index one from the back. ugh. w/e, that's fine.
card_names_file.seek(eof - 1, os.SEEK_SET)

# Oh wait, .read() will advance my pointer. Oh hey Python 3 doesn't let me
# use .peek() either. Fantastic. I'll have to read this...
if card_names_file.read() == ',':
    # Then go back to where I was by indexing from front AGAIN
    card_names_file.seek(eof - 1, os.SEEK_SET)
    # Then remove last character.
    card_names_file.truncate()

这是我几乎见过的最愚蠢的代码，到目前为止我已经花了 2 个半小时试图从文件后面删除一个字符，这看起来像是一个 hack。

另一种方法是我的代码看起来像这样

# open file
with open(file, a+)
    # do stuff

# open same file
with open(file, w+b)
    # do different stuff

但我实际上也无法让它工作。

Answer 1

底层缓冲区确实有一个您正在寻找的peek() 方法，所以：

f = open('FILE', 'a+')
f.seek(f.seek(0, os.SEEK_END) - 1)
# or with the same effect you can also:
os.lseek(f.fileno(), -1, os.SEEK_END)
# Actually in append mode we could just seek by -1 from where we are
# (cursor at the end after opening)
f.tell()  # just to see...
f.buffer.peek(1)
f.tell()  # ...still where we were

或者，您也可以使用os.pread()。例如：

os.pread(f.fileno(), 1, os.fstat(f.fileno()).st_size - 1)

这并不像依赖更高级别的抽象访问文件那样惯用，但我会invoke：“虽然实用性胜过纯度。”