On Sat, 29 Jan 2011, Aahz wrote:
> In article <mailman.1412.1296196161.6505.python-l...@python.org>,
>
> John O'Hagan <resea...@johnohagan.com> wrote:
[...]
> >
> >def lastline(filename):
> > offset = 0
> > line = ''
> > with open(filename) as f:
> > while True:
> > offset -= 1
> > f.seek(offset, 2)
> > nextline = f.next()
> > if nextline == '\n' and line.strip():
> > return line
> > else:
> > line = nextline
>
> It's a Bad Idea to mix direct file operations with the iterator API.
I didn't know that; from the docs on file objects:
"As a consequence of using a read-ahead buffer, combining next() with other
file
methods (like readline()) does not work right. However, using seek() to
reposition the file to an absolute position will flush the read-ahead buffer."
You are right in general, but the second sentence is why I got away with it in
this case.
> Use f.read() instead of f.next().
Which actually ends up improving the code as well:
def lastline(filename):
offset = 0
with open(filename) as f:
while 1:
f.seek(offset, 2)
if f.tell() == 0:
return f.read().strip()
line = f.read()
if line.strip() and line[0] == '\n':
return line.strip()
offset -= 1
although Tim Chase's solution covers files with very long lines.
John
--
http://mail.python.org/mailman/listinfo/python-list
