Hi Cam,
Please submit the modified version, and, if you want, create a separate ticket
for "seqable?". I would like to review the latter separately.
Stu
> Hi again, I modified my-flatten to return the empty list for sets and
> maps as core/flatten does. It doesn't seem to handle Arrays anymore
> though. I'm assuming it's because ArrayList and (int-array ...) don't
> implement Sequential. None the less should I still submit this
> modified version that behaves just like core/flatten?
>
> (defn my-flatten
> [coll]
> (lazy-seq
> (when-let [coll (if (sequential? coll) (seq coll))]
> (let [x (first coll)]
> (if (sequential? x)
> (concat (my-flatten x) (my-flatten (next coll)))
> (cons x (my-flatten (next coll))))))))
>
> Might it be worth promoting "seqable?" to core? In that case flatten
> would handle pretty much everything you could throw at it like you'd
> expect. I don't speak for everyone but when I saw sequential? I
> assumed it would have the semantics that seqable? does.
>
>
> On Jul 13, 11:04 am, Cam <dlocpuw...@gmail.com> wrote:
>> Hi Stuart,
>>
>> Thanks for checking that out for me! Sorry for not realizing in the
>> first place.
>>
>> I of course would be happy to submit a patch. Should I submit that
>> here or over on the assembla page?
>>
>> On Jul 13, 9:10 am, Stuart Halloway <stuart.hallo...@gmail.com> wrote:
>>
>>
>>
>>> Hi Cam,
>>
>>> Your tests aren't testing the interesting part without a doall.
>>
>>> That said, my quick tests with doall show your approach faring even better.
>>> :-) Also, I think what my-flatten does with Java arrays is intuitive (and
>>> the current flatten not so much).
>>
>>> A patch that preserves the semantics of the existing flatten (except for
>>> working with Java arrays) would be welcome.
>>
>>> Thanks!
>>> Stu
>>
>>>> Another flatten thread! Sorry..
>>
>>>> Hello all, before I realized there was a flatten in the master branch
>>>> (and before I looked at contrib) I wrote this pretty standard code:
>>
>>>> (defn my-flatten [coll]
>>>> (lazy-seq
>>>> (when-let [coll (seq coll)]
>>>> (let [x (first coll)]
>>>> (if (sequential? x)
>>>> (concat (my-flatten x) (my-flatten (next coll)))
>>>> (cons x (my-flatten (next coll))))))))
>>
>>>> (There's very similar versions on the boards. I'm not claiming this is
>>>> anything amazing or unique.)
>>
>>>> It's not as elegant as what's in core, but in my micro benchmarks (ran
>>>> on my laptop; 2.26 core 2 and 4gb ram) it seems to perform a bit
>>>> better, _especially_ in the already flattened case. It behaves just
>>>> like core/flatten except that it doesn't return an empty list when
>>>> passed a map or set, it just returns whatever you gave it but with the
>>>> top level converted to a seq. I'm pretty much a clojure noob, so are
>>>> there any hidden detractors of this implementation as opposed to the
>>>> version introduced in 1.2?
>>
>>>> Also, quick note, if you swap the call to sequential? with seqable?
>>>> from contrib/core, it flattens maps and sets like you'd expect as
>>>> well.
>>>> Here is how it looks
>>>> user=> (my-flatten #{1 {2 3} 4 [5 6 7 #{8 {9 10}}]})
>>>> (1 2 3 4 5 6 7 9 10 8)
>>
>>>> And for the micro-benchmarks (using "sequential?"):
>>
>>>> user=> (time (dotimes [_ 1e7] (flatten [1 2 3 4])))
>>>> "Elapsed time: 14,661.592 msecs"
>>>> nil
>>
>>>> user=> (time (dotimes [_ 1e7] (my-flatten [1 2 3 4])))
>>>> "Elapsed time: 922.268 msecs"
>>>> nil
>>
>>>> user=> (time (dotimes [_ 1e7] (flatten [1 [2 [3 [4 [5 [6 [7 [8]
>>>> [[[9]]] 10 [11] 12 [13 14 [15]]]]]]]]])))
>>>> "Elapsed time: 18,147.959 msecs"
>>>> nil
>>
>>>> user=> (time (dotimes [_ 1e7] (my-flatten [1 [2 [3 [4 [5 [6 [7 [8]
>>>> [[[9]]] 10 [11] 12 [13 14 [15]]]]]]]]])))
>>>> "Elapsed time: 6,088.914 msecs"
>>>> nil
>>
>>>> user=> (time (dotimes [_ 1e7] (flatten [[1 2 3 4 5 6 7 8 9 10]])))
>>>> "Elapsed time: 11,696.693 msecs"
>>>> nil
>>
>>>> user=> (time (dotimes [_ 1e7] (my-flatten [[1 2 3 4 5 6 7 8 9 10]])))
>>>> "Elapsed time: 1,533.983 msecs"
>>>> nil
>>
>>>> Thoughts?
>>
>>>> --
>>>> 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
