I used "take 0" as a simple example to illustrate the problem. But in general the standard mapcat evaluates terms than are needed. My f function does a web call and processes the JSON records produced by this call, so each extra call is a significant performance and resource hit.
On Aug 24, 2011, at 10:57 AM, Steve Miner <stevemi...@gmail.com> wrote: > I'd be careful about writing my own version of mapcat. You might save a call > or two on the f function, but you'll probably lose in the general. Try some > experiments with larger take numbers. You might be better off concentrating > on your f function. I've found that trying to special-case degenerate cases > (for empty collections, etc.) has actually hurt my performance, presumably > because my cleverness got in the way of hotspot optimizations. > > > On Aug 22, 2011, at 10:32 AM, Asim Jalis wrote: > >> user=> (defn f [[x]] (println "computing x:" (inc x)) (vector (inc x))) >> #'user/f >> user=> (->> (iterate f [0]) (take 0)) >> () >> user=> (->> (iterate f [0]) (apply concat) (take 0)) >> computing x: 1 >> computing x: 2 >> computing x: 3 >> () >> user=> (->> (iterate f [0]) (mapcat identity) (take 0)) >> computing x: 1 >> computing x: 2 >> computing x: 3 >> () >> >> Is there a way to rewrite mapcat (or apply concat) so that they don't >> evaluate the incoming seq unnecessarily? >> >> -- >> 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 > > -- > 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 -- 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
