hadley wickham wrote: > I read this to mean that setting a value in list will be O(n) (where n > >is the length of the new list) - If you blindly expect a list to act >like a hash table you will be badly surprised. If you are copying an >algorithm from another language, you will probably need to rethink the >way that updates work. If however, you just want an object that can >be indexed by a string, then a list will be fine. >
I would think that when translating from another language, it is best to write it in R in the simplest way, which probably means using a list. Then if it turns out to be too slow, try doing something fancy. I suspect that speed improvements are seldom necessary -- I can't believe how fast most computations are these days. I don't see that worrying about O(1) versus O(n) is likely to get you very far. Just try it out on a normal size problem, and try it out on a very large problem and see if it is good enough. Patrick Burns [EMAIL PROTECTED] +44 (0)20 8525 0696 http://www.burns-stat.com (home of S Poetry and "A Guide for the Unwilling S User") > > > > > ______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
