Re: Parameter ordering on map/reduce
Hi Nathan, There is actually a simple answer to your question. map can take a fuction of multiple parameters along with multiple collections. i.e. (defn f [x y] (+ x y)) (map f [1 2] [3 4]) = (4 6) (Warning I did this computation in the Clojure instance in my head, so some details may be slightly off.) Paul On 3/31/09, Nathan Sorenson n...@sfu.ca wrote: First of all, I would like to thank Rich and this community for producing such a pleasurable language, and for putting up with with all the unpleasant nit-picking of new users. That being said... I am curious as to why the function parameter is specified before the collection parameter in map/reduce. I have never used a lisp before, and may not be aware of idiomatic style, but it seems to be the convention elsewhere in Clojure (assoc, conj, .method calls, etc...) to have the altered data structure be the first parameter. Would this not allow mixing map into a threaded expression: (- [1 2 3] (map inc) (assoc 0 4) (reduce +) (conj :anotherthing)) Perhaps this style is rare in practice? Certainly it is easy enough to write a custom map/reduce which acts this way, so I suppose my question is mostly philosophical. --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
Parameter ordering on map/reduce
First of all, I would like to thank Rich and this community for producing such a pleasurable language, and for putting up with with all the unpleasant nit-picking of new users. That being said... I am curious as to why the function parameter is specified before the collection parameter in map/reduce. I have never used a lisp before, and may not be aware of idiomatic style, but it seems to be the convention elsewhere in Clojure (assoc, conj, .method calls, etc...) to have the altered data structure be the first parameter. Would this not allow mixing map into a threaded expression: (- [1 2 3] (map inc) (assoc 0 4) (reduce +) (conj :anotherthing)) Perhaps this style is rare in practice? Certainly it is easy enough to write a custom map/reduce which acts this way, so I suppose my question is mostly philosophical. --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
Re: Parameter ordering on map/reduce
If I remember correctly (I wasn't able to find the thread where this same subject was explained), one must make a difference between functions intended to work on data structures, and functions intended to work on sequences. Data structures are not volatile as sequences are, and those functions that manipulate them generally take the data structure (vector, map, set, etc.) as their first argument. It makes sense to create a partial function with the data structure captured by the partial call. Sequences are volatile, and thus less subject to be captured by a partial call. And to enforce the difference between the 2, sequences are generally the 2d or last argument to such functions. HTH, -- Laurent 2009/3/31 Nathan Sorenson n...@sfu.ca First of all, I would like to thank Rich and this community for producing such a pleasurable language, and for putting up with with all the unpleasant nit-picking of new users. That being said... I am curious as to why the function parameter is specified before the collection parameter in map/reduce. I have never used a lisp before, and may not be aware of idiomatic style, but it seems to be the convention elsewhere in Clojure (assoc, conj, .method calls, etc...) to have the altered data structure be the first parameter. Would this not allow mixing map into a threaded expression: (- [1 2 3] (map inc) (assoc 0 4) (reduce +) (conj :anotherthing)) Perhaps this style is rare in practice? Certainly it is easy enough to write a custom map/reduce which acts this way, so I suppose my question is mostly philosophical. --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
Re: Parameter ordering on map/reduce
Hi Nathan, On 31 Mrz., 09:40, Nathan Sorenson n...@sfu.ca wrote: First of all, I would like to thank Rich and this community for producing such a pleasurable language, and for putting up with with all the unpleasant nit-picking of new users. That being said... I am curious as to why the function parameter is specified before the collection parameter in map/reduce. I have never used a lisp before, and may not be aware of idiomatic style, but it seems to be the convention elsewhere in Clojure (assoc, conj, .method calls, etc...) to have the altered data structure be the first parameter. I believe this is to allow more than one sequence in map expressions: (map - '(1 2 3) '(3 2 1)) = (-2 0 2) and for reduce, it takes another optional arg to start the reduce with: (reduce conj [1 2] '(a b)) = [1 2 a b] Would this not allow mixing map into a threaded expression: (- [1 2 3] (map inc) (assoc 0 4) (reduce +) (conj :anotherthing)) Perhaps this style is rare in practice? Certainly it is easy enough to write a custom map/reduce which acts this way, so I suppose my question is mostly philosophical. there was a thread devoted to a macro to express this kind of piping, http://groups.google.com/group/clojure/browse_thread/thread/66ff0b89229be894/e78c3585589cdbf6?lnk=gstq=pipe#e78c3585589cdbf6 (defmacro pipe Threads the expr through the forms. Inserts x as the last item in the first form, making a list of it if it is not a list already. If there are more forms, inserts the first form as the last item in second form, etc. ([x form] (if (seq? form) `(~(first form) ~@(rest form) ~x) (list form x))) ([x form more] `(pipe (pipe ~x ~form) ~...@more))) which is the same as the - except it threads its first arg to the last position of each form instead of the second (pipe x (f a b)) = (f a b x) Using this macro is IMO good for expressing long chains of filtering, mapping and reducing, its more readable, less nested and easier to edit than the equivalent nested expression. I have some code which is using this macro, but most of the time I just don't need such long pipes or I'm using a let to create named intermediate results for debugging and sanity. In the above mentioned thread, there is also a more general let- macro which would make your example work without any modifications to the core functions: (let- x [1 2 3] (map inc x) (into [] x) (assoc x 0 4) (reduce + x) (list x) (conj x :anotherthing)) = (:anotherthing 11) erik --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
Re: Parameter ordering on map/reduce
Would this not allow mixing map into a threaded expression: You might also be interested in this thread: http://groups.google.com/group/clojure/browse_thread/thread/66ff0b89229be894 Where various people created a let- (or pipe) macro for such a purpose. --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
Re: Parameter ordering on map/reduce
I think clojure is a little torn between functional thinking and oo thinking - java vs haskell (i must admit that I am not originally a lisper, more of a java and functional guy) Functional / haskell style is what you see in map. that collection is at the end. The preceding arguments are like parameters for the function, suited for partial invocations map f coll take 10 coll java (or maybe oo) style is to think abvout the object as the first parameter conj coll e nth coll 10 I think it is not so intuitive If the function exist in the functional world (reduce filter map cons..) then the collection is at the end, otherwise it tends to be the first parameter. But, of course, if the function has a variable length argument list, that must be at the end --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
Re: Parameter ordering on map/reduce
The let- macro looks very useful. Thank-you! --~--~-~--~~~---~--~~ 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 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 -~--~~~~--~~--~--~---
