Question

我需要下载文本文件的zip存档，将存档中的每个文本文件发送给其他处理程序进行处理，最后将解压缩的文本文件写入磁盘。

我有以下代码。它在同一个文件上使用多个打开/关闭，这看起来并不优雅。如何让它更优雅高效？

zipped = urllib.urlopen('www.abc.com/xyz.zip')
buf = cStringIO.StringIO(zipped.read())
zipped.close()
unzipped = zipfile.ZipFile(buf, 'r')
for f_info in unzipped.infolist():
   logfile = unzipped.open(f_info)
   handler1(logfile)
   logfile.close()   ## Cannot seek(0). The file like obj does not support seek()
   logfile = unzipped.open(f_info)
   handler2(logfile)
   logfile.close()
   unzipped.extract(f_info)

Answer 1

您的答案在您的示例代码中。只需使用StringIO缓冲日志文件：

zipped = urllib.urlopen('www.abc.com/xyz.zip')
buf = cStringIO.StringIO(zipped.read())
zipped.close()
unzipped = zipfile.ZipFile(buf, 'r')
for f_info in unzipped.infolist():
   logfile = unzipped.open(f_info)
   # Here's where we buffer:
   logbuffer = cStringIO.StringIO(logfile.read())
   logfile.close()

   for handler in [handler1, handler2]:
      handler(logbuffer)
      # StringIO objects support seek():
      logbuffer.seek(0)

   unzipped.extract(f_info)

Answer 2

你可以这样说：

handler_dispatch(logfile)

和

def handler_dispatch(file):
   for line in file:
      handler1(line)
      handler2(line)

甚至通过构造具有多个handlerN函数的Handler类并在handler_dispatch内应用它们来使其更具动态性。像

class Handler:
    def __init__(self:)
        self.handlers = []

  def add_handler(handler):
      self.handlers.append(handler)

  def handler_dispatch(self, file):
      for line in file:
          for handler in self.handlers:
              handler.handle(line)

Answer 3

打开zip文件一次，遍历所有名称，提取每个名称的文件并进行处理，然后将其写入磁盘。

像这样：

for f_info in unzipped.info_list():
    file = unzipped.open(f_info)
    data = file.read()
    # If you need a file like object, wrap it in a cStringIO
    fobj = cStringIO.StringIO(data)
    handler1(fobj)
    handler2(fobj)
    with open(filename,"w") as fp:
        fp.write(data)

你明白了

在Python中多次读取同一文件

3 个答案: