If the sequence is already realized, or is cheap, and you want only a very small random subset of it, you can do better than shuffling the whole thing. Fliebel and I played around with several solutions to this, some time ago. I can't find the whole thing, but some interesting examples and benchmarking data are at https://gist.github.com/805747 if you want to try it out.
On Jul 11, 3:57 am, Tassilo Horn <tass...@member.fsf.org> wrote: > Sunil S Nandihalli <sunil.nandiha...@gmail.com> writes: > > Hi Sunil, > > > I think a lazy version of shuffle would be a good addtion to the > > core. I tried googling for one and found > >http://ytakenaka.blogspot.com/2011/05/lazy-shuffle-clojure.htmlwhich > > was infact slower than original shuffle and was unable to reproduce > > the performance claims made there in. > > I think there is no way to implement a "good" lazy version of shuffle. > Correct shuffling, i.e., the result should be a real pseudo-random > permutation of all elements, can only be defined on finite seqs anyway. > And there, one has to realize the given seq. In the code you cite, > there's > > (defn lazy-shuffle [lst fetch-len] > (let [len (count lst) > ^^^^^^^^^^^ > > which does exactly that. So here only the cheap creation of the > shuffled seq is lazy, but the possibly expensive calculation of the > given seq is (and has to be done) directly. > > Bye, > Tassilo -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
