To continue my previous email, here is the implementation of map which is
both reducible and (lazy) seqable:
(defn my-map
[f c]
(let [coll (clojure.core.reducers/map f c)
d (delay (clojure.core/map f c))]
(reify
clojure.core.protocols/CollReduce
(coll-reduce [_ f1]
(clojure.core.protocols/coll-reduce coll f1))
(coll-reduce [_ f1 init]
(clojure.core.protocols/coll-reduce coll f1 init))
clojure.core.reducers/CollFold
(coll-fold [_ n combinef reducef]
(clojure.core.reducers/coll-fold coll n combinef reducef))
clojure.lang.Seqable
(seq [_] @d))))
user> (def x (my-map inc (range 100)))
#'user/x
user> x
#<user$my_map$reify__1366 user$my_map$reify__1366@300f8aa3>
user> (seq x) ;; to lazy sequence
(1 2 3 4 5 6 7 8 9 10 11 12 . . . 100)
user> (last x) ;; using sequence
100
user> (reduce + x) ;; using reducer
5050
user> (clojure.core.reducers/fold + x) ;; using folder
5050
JW
On Thursday, August 8, 2013 4:46:37 PM UTC+2, Jozef Wagner wrote:
>
> Wow, thank you very much! A perfect solution.
>
> At the end, wouldn't be good if the reducers would implement alongside
> CollReduce also a Seqable interface, so that reducers could be used as a
> drop in replacement for today's sequence functions (map, filter, ...)?
> CollReduce implements 'eager' computations (when calling reduce and fold)
> while the Seqable would implement a lazy ones (when calling seq).
>
> Functions working with seqs (e.g. first, rest) always call seq before
> working with the collection, so this would be transparent for them. It
> may be a breaking change and complects things a bit though...
>
> JW
>
> On Thursday, August 8, 2013 3:34:31 PM UTC+2, Christophe Grand wrote:
>>
>> ArrayDeque based versions:
>>
>> (defn drop-last [n coll]
>> (reducer coll
>> (fn [f1]
>> (let [buffer (java.util.ArrayDeque. (int n))]
>> (fn self
>> ([] (f1))
>> ([ret x]
>> (.add buffer x)
>> (if (<= (count buffer) n)
>> ret
>> (f1 ret (.pop buffer)))))))))
>>
>> (defn take-last [n coll]
>> (reify clojure.core.protocols.CollReduce
>> (coll-reduce [this f1]
>> (clojure.core.protocols/coll-reduce this f1 (f1)))
>> (coll-reduce [_ f1 init]
>> (clojure.core.protocols/coll-reduce
>> (doto (clojure.core.protocols/coll-reduce
>> coll
>> (fn [^java.util.Deque q x]
>> (.add q x)
>> (when (> (count q) n)
>> (.pop q))
>> q) (java.util.ArrayDeque. (int n))) prn)
>> f1 init))))
>>
>>
>>
>> On Thu, Aug 8, 2013 at 3:16 PM, Christophe Grand <[email protected]>wrote:
>>
>>> You need to use a buffer to defer calls to the reduced function
>>>
>>> (defn drop-last [n coll]
>>> (reducer coll
>>> (fn [f1]
>>> (let [buffer
>>> (atom clojure.lang.PersistentQueue/EMPTY)]
>>> (fn self
>>> ([] (f1))
>>> ([ret x]
>>> (let [b (swap! buffer conj x)]
>>> (if (<= (count @buffer) n)
>>> ret
>>> (do
>>> (swap! buffer pop)
>>> (f1 ret (peek b)))))))))))
>>>
>>> An array or a ring buffer should be used instead of the atom and
>>> persistent queue combo to reduce allocation.
>>>
>>> take-last is harder because you can't know when the reduction is over
>>> when using #'reducer, so you have to implement CollReduce yourself:
>>>
>>> (defn take-last [n coll]
>>> (reify clojure.core.protocols.CollReduce
>>> (coll-reduce [this f1]
>>> (clojure.core.protocols/coll-reduce this f1 (f1)))
>>> (coll-reduce [_ f1 init]
>>> (clojure.core.protocols/coll-reduce
>>> (clojure.core.protocols/coll-reduce
>>> coll
>>> (fn [q x]
>>> (let [q (conj q x)]
>>> (if (<= (count q) n)
>>> q
>>> (pop q)))) clojure.lang.PersistentQueue/EMPTY)
>>> f1 init))))
>>>
>>> again, use of a mutable array/buffer would be preferable.
>>>
>>> hth,
>>>
>>> Christophe
>>>
>>> On Thu, Aug 8, 2013 at 1:00 PM, Jozef Wagner <[email protected]>wrote:
>>>
>>>> Is it possible to implement efficient butlast (and drop-last,
>>>> take-last) with reducers? The only solution I can think of needs
>>>> additional
>>>> reduce to compute count, which may often be undesirable.
>>>>
>>>> Or is it OK to say that reducers are not designed for such cases?
>>>>
>>>> JW
>>>>
>>>> --
>>>> --
>>>> You received this message because you are subscribed to the Google
>>>> Groups "Clojure" group.
>>>> To post to this group, send email to [email protected]
>>>> Note that posts from new members are moderated - please be patient with
>>>> your first post.
>>>> 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
>>>> ---
>>>> You received this message because you are subscribed to the Google
>>>> Groups "Clojure" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send
>>>> an email to [email protected].
>>>> For more options, visit https://groups.google.com/groups/opt_out.
>>>>
>>>>
>>>>
>>>
>>>
>>>
>>> --
>>> On Clojure http://clj-me.cgrand.net/
>>> Clojure Programming http://clojurebook.com
>>> Training, Consulting & Contracting http://lambdanext.eu/
>>>
>>
>>
>>
>> --
>> On Clojure http://clj-me.cgrand.net/
>> Clojure Programming http://clojurebook.com
>> Training, Consulting & Contracting http://lambdanext.eu/
>>
>
--
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to [email protected]
Note that posts from new members are moderated - please be patient with your
first post.
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
---
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
