Is reverse reading possible?

[Anthony Liu]

Jeff and Ivo,

Thank you very much for your solution.  Yes, I figure
out it is easy for an ascii text.

The files (nearly 4M each)I am gonna read are a
mixture of Chinese and English, where each Chinese
character has 2 bytes and each English (ASCII) has 1
byte, although the majority of the texts is Chinese.

So, if we reverse-read it without taking into
consideration the 2-byte Chinese characters, we are
gonna get the Chinese characters spelled out in the
wrong way - unreadable!

So, it might be tedious to read such a text reversely
unless you have a smarter idea.  Do you?



--- Jeff Epler <jepler at unpythonic.net> wrote:
> on seekable, byte-oriented binary files, sure.
> 
> import errno
> 
> SEEK_SET, SEEK_CUR, SEEK_END = range(3)
> 
> class Reversed:
>     def __init__(self, f):
>         self.f = f
>         self.f.seek(-1, SEEK_END)
>         self.at_eof = 0
> 
>     def read_byte(self):
>         if self.at_eof: return ''
>         byte = self.f.read(1)
>         try:
>             self.f.seek(-2, SEEK_CUR)
>         except IOError, detail:
>             if detail.errno == errno.EINVAL:
>                 self.at_eof = 1
>             else:
>                 raise
>         return byte
> 
>     def __iter__(self): return self
> 
>     def next(self):
>         r = self.read_byte()
>         if r == '': raise StopIteration
>         return r
> 
> >>> "".join(Reversed(file("/etc/redhat-release",
> "rb")))
> '\n)worraY( 1 esaeler eroC arodeF'
> 
> Jeff


__________________________________
Do you Yahoo!?
Yahoo! Mail - More reliable, more storage, less spam
http://mail.yahoo.com