On Sat, Mar 22, 2008 at 09:46:58AM +0100, Andreas Kostyrka wrote: > > Basically, strings are immutable. If you need to append something to a > string, you need to construct a new string object with the new value. > > Now if you are using this to collect huge outputfiles in pieces, one of > the common idioms in Python is collecting it in a list, and converting > to string at the end: > > collector = [] > for i in xrange(100000): > collector.append((str(i) * 80)[0:80]) > > string = "".join(collector) > assert len(string) == 8000000 # ~8MB
That was formerly good advice for sure, and it probably still is good advice. But, read the following note from http://docs.python.org/lib/typesseq.html: (6) If s and t are both strings, some Python implementations such as CPython can usually perform an in-place optimization for assignments of the form s=s+t or s+=t. When applicable, this optimization makes quadratic run-time much less likely. This optimization is both version and implementation dependent. For performance sensitive code, it is preferable to use the str.join() method which assures consistent linear concatenation performance across versions and implementations. Changed in version 2.4: Formerly, string concatenation never occurred in-place. As the above note says, this optimization is implementation dependent. In particular, if you plan on moving your code to Jython, then follow Andreas's suggestion to use list append followed by string join. - Dave -- Dave Kuhlman http://www.rexx.com/~dkuhlman _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor
