Reply to thread

Message: [QUOTE="Rasmus Villemoes, post: 4863125"] Actually, I _do_ know some algorithm theory, and I also know that in concrete situations with some upper bound on N an O(N log N) algorithm may be faster than an O(N) algorithm, because of the unmentioned constants hidden in the O-notation. Yes, that was what I thought. I didn't write this, but I know that my data sets will never have more than around 40 elements (most probably even shorter). But I will need to do this lots of times over, so I'm interested in using the method which is fastest for short data sets. Once I get some real data to work with, I will test the different methods you and others have suggested. No, I didn't say it was clearer. The problem with string lengths is non-existent in the real case I'm considering, because, as mentioned, I am really looking at some other properties of $h{$a}. [/QUOTE]