On Wed, Aug 27, 2008 at 12:11 AM, Chouser <[EMAIL PROTECTED]> wrote:
>
> > You could simplify your reduce fn a bit by using merge-with:
> Or even:
>
> (defn test-shuffle-uniformity
> "Shuffles a 3-item collection n times, mapping each result to the number
> of
> times it's hit. If the distribution isn't uniform we've let bias into the
> random numbers or the algorithm. Typical results on my machine:
> (2 1 3) : 16611
> (3 2 1) : 16771
> (1 3 2) : 16707
> (1 2 3) : 16766
> (3 1 2) : 16555
> (2 3 1) : 16590"
> ([n] (test-shuffle-uniformity n shuffle))
> ([n shuffle]
> (let [coll [1 2 3]
> results (apply merge-with + {}
> (take n (repeatedly (fn [] {(shuffle coll) 1}))))]
> (doseq [k v] results
> (println k ":" v)))))
>
Interesting suggestion. I think reduce expresses my intent a little better,
but I need to keep merge-with in mind for other uses.
> It'd be nice to use the #(...) syntax instead of (fn [] ...), but that
> won't work here since #({key val}) would try to call the map with no
> arguments. You'd have to say #(hash-map key val) and (fn [] {key
> val}) is arguably clearer.
>
Heh, having #( ) available gives me a concision addiction. I'm trying to
heed the advice of the docs not to replace fn everywhere, but it's always
tempting to use. I definitely tried to write #(rand) in shuffle, but ran
into the same no-arguments problem.
Also note the destructuring in doseq to save a few keystrokes.
>
Ah, I didn't realize at the time that doseq supported destructuring, but
that rocks.
> Or you can rely on the fact that vectors are comparable themselves.
> The first member is the most significant, so I'd think this would
> produce acceptable results. It's more succinct and appears to double
> the speed:
>
> (defn sort-by-mem-keys
> "Like clojure/sort-by but only calls keyfn once on each item in coll."
> ([keyfn coll]
> (let [key-vals (map (fn [x] [(keyfn x) x]) coll)]
> (map second (sort key-vals)))))
>
That's very clean. When duplicate keys arise that implementation is probably
has similar behavior to the original, assuming the values in coll are
Comparable. Either way, I haven't tested mine for uniformity with large
inputs, and given the superiority of the Java solution, I'm done.
Thanks for the pointers!
Shawn
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
-~----------~----~----~----~------~----~------~--~---
