I am trying to return from
'('(animal '(name a b c) '(age 1 2 3) '(family x y z)) only '(name age family). How do I do that? I was trying this:
(car(apply append(cons (cadr (car list)) '())))
but I get an error.
I am not sure the list you want to handle is:
'('(animal '(name a b c) '(age 1 2 3) '(family x y z)) or
'(animal (name a b c) (age 1 2 3) (family x y z))
The latter one makes more sense to me. Your (sub) problem is to return a list containing all first elements of sublists. You may need to use recursion. Something like:
(define get_list_of_first
(lambda (lol)
(if (null? lol) '()
(cons (car (car lol)) (get_list_of_first (cdr lol)))
)
)
)
This is a very basic example to learn Scheme. You need to use the above function to continue your assignment (I guess). Have fun.
Related
I have a partition for a quicksort:
(define (partition pivot lst)
((lambda (s) (s s lst list))
(lambda (s l* c)
(if (null? l*)
(c '() '())
(let ((x (car l*)))
(s s (cdr l*)
(lambda (a b)
(if (< x pivot)
(c (cons x a) b)
(c a (cons x b))))))))))
partition code source
Testing:
=>(partition '5 '(1 3 5 7 9 8 6 4 2))
;Value: ((1 3 4 2) (5 7 9 8 6))
How can I implement this partition in a quicksort? I've tried this so far:
(define (quicksort lst)
(if (null? lst) '()
(let* ((y (car lst))
(pn (partition y (cdr lst))))
(append (quicksort (car pn))
(list y)
(quicksort (cdr pn))))))
First, your code is trivially fixed by changing one cdr to cadr:
(define (partition pivot lst)
((lambda (s) (s s lst list))
......)) ; ^^^^ `list` the top continuation
(define (quicksort lst)
(if (null? lst) '()
(let* ((y (car lst))
(pn (partition y (cdr lst))))
(append
(quicksort (car pn))
(list y)
(quicksort (cadr pn))))))
;; ^^^^ cdr --> cadr
because the top continuation used in partition is list, and so the call
(partition pivot lst)
is equivalent to the call
(list { x IN lst SUCH THAT x < pivot }
{ x IN lst SUCH THAT x >= pivot } )
(the parts in {...} are pseudocode, where we don't care about the implementation, just the results)
And so to access the two parts of that list built by partition you need to use car and cadr.
Or you could keep the cdr in the accessing part of your code in quicksort if you'd change that top continuation to cons:
(define (partition pivot lst)
((lambda (s) (s s lst cons))
......)) ; ^^^^ `cons` the top continuation
(define (quicksort lst)
(if (null? lst)
'()
(let* ((y (car lst))
(pn (partition y (cdr lst))))
(append
(quicksort (car pn))
(list y)
(quicksort (cdr pn))))))
;; ^^^^ `cdr` works fine with `cons`
This because of the general principle in programming, where the functions used to build our data dictate which functions are to be used to access that data:
(list <A> <B> )
car cadr
(cons <A> <B> )
car cdr
( this particular correspondence is because (list <A> <B>) is the same as (cons <A> (cons <B> '())) and (cadr <C>) is the same as (car (cdr <C>)): )
(list <A> <B> )
=
(cons <A> (cons <B> '()))
car cdr
car
And conversely, the functions we use to access our data dictate the implementation of the function which must be used to build that data.
Of course that way of coding in your question is considered unnecessarily convoluted by modern standards since it emulates recursion through argument passing and reuse, -- just like in the famous Y combinator, -- but any Scheme worthy of its name already supports recursion.
So this partition would normally be written as the fully equivalent yet more readable code using the "named let" construct,
(define (partition pivot lst)
(let s ( (l* lst) ; first `l*` is `lst`
(c cons) ) ; top `c` is `cons`
(if (null? l*)
(c '() '())
(let ((x (car l*)))
(s (cdr l*)
(lambda (a b)
(if (< x pivot)
(c (cons x a) b)
(c a (cons x b)))))))))
except the name loop is conventionally used in place of s here (which itself most probably is intended as the shortening of "self").
But the true trouble with your quicksort/partition pair is algorithmic.
Yes I say pair (in non-cons sense of course) since the two go together -- just as with the data access/creation functions which must work together too.
Implementation of one dictates the implementation of the other -- in both directions, too. partition's code dictates quicksort's, or if we'd written quicksort first, we'd need to implement the partition in the corresponding way -- so that the two work together. Which means quicksort indeed producing the correct results, turning any input list into a sorted one:
(quicksort lst) --->
{ xs SUCH THAT
FOR ANY splitting xs = { ..., x, ...ys }
AND ANY splitting ys = { ..., y, ... }
IT HOLDS THAT x <= y
AND ALSO xs is a permutation of lst
(which implies (length lst) == (length xs))
}
So then, what is that trouble? It is that the true quicksort does no work whatsoever after the partitioning. None:
(define (quicksort! lst)
(if (null? lst)
'()
(let* ((y (car lst))
(pn (partition! y lst)))
(quicksort! (car pn)) ; no `append`, NB!
(quicksort! (cdr pn))))) ; no (list y) either
How is that even possible? What kind of partition! implementation would make that work? Well, most certainly not a functional one.
Instead it must be changing (i.e. mutating) the very lst itself somehow:
{ a, b, c, ....., k, l, m, ..... }
-->
{ d, e, ...., p, n, o, ..... }
~~~~~~~~~~~ ~~~~~~~~~~~
where we denote with p the partition point -- so that indeed all that's left to do after this kind of partitioning "in-place" is to sort the first part, and then to sort the second part, -- and then there's nothing more left to be done, after that! Which was the key insight in the original Tony Hoare's formulation of it:
TO SORT
{ a, b, c, ....., k, l, m, ..... } DO:
PARTITION it into
{ d, e, ...., p, n, o, ..... } AND THEN:
~~~~~~~~~~~ ~~~~~~~~~~~
SORT! SORT!
DONE.
This partitioning is usually implemented with swap! which actually swaps two elements in the underlying data structure. Most usually that data structure is an array with its facilities to change the value stored in it at any given index.
But it can also be a list, where the change i.e. mutation can be done with the set-car! primitive.
Looks like we'd need to build a list of cdrs out of the input list, and another one in reverse, -- to be able to iterate over them in both directions, back and forth, -- to make that happen.
I'll leave that for another day, for now.
Once you have the partition, there is still a small step to do.
Take care, you need to be sure partition splits the input in smaller sets all the time. In other word, partition not to return some empty set. The pivot can go in any of the sets and use this fact to check that you do not return an empty set, in case your comparison operator does not really decrease the size of the input. This is why I inserted the equality operator -- to be able to check if I insert the pivot in the first returned set or in the second one.
(define (partition pivot lst ret)
((lambda (s)
(s s lst
(lambda (a b p*)
(if (and (null? a) (null? b))
(ret (list pivot) (cdr p*))
(if (null? a)
(ret p* b)
(if (null? b)
(ret a p*)
(if (< (car b) pivot)
(ret a (append p* b))
(if (< (car a) pivot)
(ret (append a p*) b)
(error "never here")))))))))
(lambda (s l* c)
(if (null? l*)
(c '() '() '())
(let ((x (car l*)))
(s s (cdr l*)
(lambda (a b p*)
(if (= x pivot)
(c a b (cons pivot p*))
(if (< x pivot)
(c (cons x a) b p*)
(c a (cons x b) p*))))))))))
(define choose-pivot car)
In a real implementation, you will all the time use vectors and this is why the append will not be present, as, sorting on the place, at the end of partition, both sides will be sorted relatively one to the other. Here, we need to reassemble the 2 sides using append:
(define (quicksort lst)
(if (null? lst) '()
(if (null? (cdr lst))
lst
(let* ((pivot (choose-pivot lst)))
(partition pivot lst
(lambda (p< p>)
(append
(quicksort p<)
(quicksort p>))))))))
A test:
1 ]=> (quicksort '(1 3 5 7 9 8 6 4 2))
;Value: (1 2 3 4 5 6 7 8 9)
1 ]=> (quicksort '(1 9 3 8 5 7 7 6 9 5 8 4 6 3 4 2 2 1))
;Value: (1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9)
I used as pivot the first element of the input to split, but you can redefine the choose-pivot to select other element.
In practice, this algorithm is used in combination with other sorts -- when the input has fewer than 4-8 elements, the quicksort is not recurred any more, but other sorting is used for the lowest cases of recurrence relation.
I used directly < in the code -- you can insert it as a parameter in case you prefer a more generic procedure... In any case, the operator that you use needs to simulate the equality and different of in the same time.
UPDATE I have updated the partition, such that to consider duplicated elements. In my first version, it ignored duplicated elements.
The code works like this, I pass several lists and it returns me all lists in just one. What I want it to do is that after joining the elements in a list, it removes the repeating elements. To be clearer:
The code is working like this:
> (koo '(p x k c l) '(l x y c) '(x k))
'(p x k c l l x y c x k)
I want him to come back to me like this:
> (koo '(p x k c l) '(l x y c) '(x k))
'(p y )
Here the code:
(define (koo . c)
(if (null? c)
empty
(concatenate1 (first c)
(apply xor* (rest c)))))
(define (concatenate1 l1 l2)
(if (null? l1)
l2
(cons (first l1) (concatenate1 (rest l1) l2))))
A not very efficient version using just basic racket functions:
(define (unique-elements lst)
(let ((dup (check-duplicates lst)))
(if dup
(unique-elements (remove* (list dup) lst))
lst)))
(define (xor* . lol) (unique-elements (append* lol)))
Looks for the first duplicate element, and if any, removes all instances of that element from the list, and repeat until there are no duplicates.
Documentation for check-duplicates.
Documentation for remove*
Documentation for append*
I started to learn Racket and I don't know how to check if a list is found in another list. Something like (member x (list 1 2 3 x 4 5)), but I want that "x" to be a a sequence of numbers.
I know how to implement recursive, but I would like to know if it exists a more direct operator.
For example I want to know if (list 3 4 5) is found in (list 1 2 3 4 5 6 )
I would take a look at this Racket Object interface and the (is-a? v type) -> boolean seems to be what you are looking for?, simply use it while looping to catch any results that are of a given type and do whatever with them
you may also want to look into (subclass? c cls) -> boolean from the same link, if you want to catch all List types in one go
should there be a possiblity of having a list inside a list, that was already inside a list(1,2,(3,4,(5,6))) i'm afraid that recursion is probally the best solution though, since given there is a possibility of an infinit amount of loops, it is just better to run the recursion on a list everytime you locate a new list in the original list, that way any given number of subList will still be processed
You want to search for succeeding elements in a list:
(define (subseq needle haystack)
(let loop ((index 0)
(cur-needle needle)
(haystack haystack))
(cond ((null? cur-needle) index)
((null? haystack) #f)
((and (equal? (car cur-needle) (car haystack))
(loop index (cdr cur-needle) (cdr haystack)))) ; NB no consequence
(else (loop (add1 index) needle (cdr haystack))))))
This evaluates to the index where the elements of needle is first found in the haystack or #f if it isn't.
You can use regexp-match to check if pattern is a substring of another string by converting both lists of numbers to strings, and comparing them, as such:
(define (member? x lst)
(define (f lst)
(foldr string-append "" (map number->string lst)))
(if (regexp-match (f x) (f lst)) #t #f))
f converts lst (a list of numbers) to a string. regexp-match checks if (f x) is a pattern that appears in (f lst).
For example,
> (member? (list 3 4 5) (list 1 2 3 4 5 6 7))
#t
One can also use some string functions to join the lists and compare them (recursion is needed):
(define (list-in-list l L)
(define (fn ll)
(string-join (map number->string ll))) ; Function to create a string out of list of numbers;
(define ss (fn l)) ; Convert smaller list to string;
(let loop ((L L)) ; Set up recursion and initial value;
(cond
[(empty? L) #f] ; If end of list reached, pattern is not present;
[(string-prefix? (fn L) ss) #t] ; Compare if initial part of main list is same as test list;
[else (loop (rest L))]))) ; If not, loop with first item of list removed;
Testing:
(list-in-list (list 3 4 5) (list 1 2 3 4 5 6 ))
Output:
#t
straight from the Racket documentation:
(member v lst [is-equal?]) → (or/c list? #f)
v : any/c
lst : list?
is-equal? : (any/c any/c -> any/c) = equal?
Locates the first element of lst that is equal? to v. If such an element exists, the tail of lst starting with that element is returned. Otherwise, the result is #f.
Or in your case:
(member '(3 4 5) (list 1 2 3 4 5 6 7))
where x is '(3 4 5) or (list 3 4 5) or (cons 3 4 5)
it will return '(3 4 5 6 7) if x ( searched list '(3 4 5) ) was found in the list or false (#f) if it was not found
or you can use assoc to check if your x is met in one of many lists, or :
(assoc x (list (list 1 2) (list 3 4) (list x 6)))
will return :
'(x 6)
There are also lambda constructions but I will not go in depth since I am not very familiar with Racket yet. Hope this helps :)
EDIT: if member gives you different results than what you expect try using memq instead
I'm working on this procedure which is supposed to return a list of pairs from 2 given lists. So for example (pairs '(1 2 3) '(a b c)) should return '((1.a) (2.b) (3.c)).
This is my logic so far. I would take the first element of each list and recursively call the procedure again with cdr as the new arguments. My result is returning a list such as this: (1 a 2 b 3 c)
Where is my logic going wrong? I know there is a list missing somewhere, but I'm not an expert at Scheme.
Any suggestions?
(define pairs
(lambda (x y)
(if (or (null? x) (null? y))
'()
(cons (car x)
(cons (car y)
(pairs (cdr x)(cdr y)))))))
(pairs '(1 2 3) '(a b c))
Notice that you produce a value that prints as (1 . 3) by evaluating (cons 1 3). However in your program you are doing (cons 1 (cons 3 ...)) which will prepend 1 and 3 to the following list.
In other words: Instead of (cons (car x) (cons (car y) (pairs ...))
use (cons (cons (car x) (car y) (pairs ...)).
Using map simplifies it a lot:
(define (pairs x y)
(map (λ (i j) (list i j)) x y))
Testing:
(pairs '(1 2 3) '(a b c))
Output:
'((1 a) (2 b) (3 c))
The result you're looking for should look like this:
((1 a) (2 b) (3 c))
In reality this structure is similar to this:
(cons
(cons 1 a)
(cons
(cons 2 b)
(cons
(cons 3 c)
'()
)
)
)
So what you're looking for is to append pairs to a list instead of adding all items to the list like you do. Simply your result looks like this:
(1 (2 (pairs ...)))
Your code should look like this:
(define pairs
(lambda (x y)
(if (or (null? x) (null? y))
'()
(cons
(cons (car x) (car y))
(pairs (cdr x) (cdr y))))))
This code might work, but it isn't perfect. We could make the code pass the list we create as a third parameter to make the function tail recursive.
You'd have something like this:
(define pairs
(lambda (x y)
(let next ((x x) (y y) (lst '()))
(if (or (null? x) (null? y))
(reverse lst)
(next (cdr x)
(cdr y)
(cons
(cons (car x) (car y))
lst))))))
As you can see, here since we're adding next element at the beginning of the list, we have to reverse the lst at the end. The difference here is that every time next is called, there is no need to keep each state of x and y in memory. When the named let will return, it won't be necessary to pop all the values back to where it called. It will simply return the reversed list.
That said, instead of using reverse we could simply return lst and use (append lst (cons (car x) (car y))) which would append the pair at the end of the list... Since lists are linked lists... in order to append something at the end of the list, scheme has to walk over all list items... which migth not be good with big list. So the solution is to add everything and at the end reorder the list as you wish. The reverse operation would happen only once.
I notice that almost all scheme functions can only return one list as output.
In the following, I would like to return multiple values of all the adjacent nodes of neighbors.
(define (neighbors l w)
(if (and (= 1 l) (= 1 w))
(list (and (l (+ 1 w))) (and (+ 1 l) w)))) ; how to output 2 or more values?
In this case I'm first testing if the node is at corner, if so, return 2 values of the coordinates where (l and w+1), (l+1 and w) basically if I'm at (1,1) return me (1,2) and (2,1)
Same applies when the node has only 1 neighbor near the edge, in this case I will have 3 values.
When no edge is nearby I will have 4 return values.
I tried to use cons, append, list, display, write none of them seems working with additional values. I need this as a sub-function of this question. How should I implement it so I could pass on the return value and use it recursively to return me all the adjacent nodes?
Edit: I found the answer: use the keyword "values" to return multiple values. Example:
(define (store l w)
(values (write l)
(write w)
(newline)
(list (+ 1 w) l)
(list w (+ 1 l))))
values, continuation passing style, and list are at least three ways of returning multiple values:
(import (rnrs))
; let-values + values
(define (foo1)
(values 1 2 3))
(let-values (((a b c) (foo1)))
(display (list a b c))
(newline))
; cps
(define (foo2 k)
(k 1 2 3))
(foo2 (lambda (a b c)
(display (list a b c))
(newline)))
; list
(define (foo3)
(list 1 2 3))
(let ((result (foo3)))
(display result)
(newline))
The Guile implementation of Scheme has a receive syntax, which it says is "much more convenient" than values. I haven't used it yet, however, but this may be useful:
http://www.gnu.org/software/guile/manual/html_node/Multiple-Values.html
You can return a pair of values in a cons cell:
(define (foo)
(cons 'a 5))
(let* ((r (foo))
(x (car r))
(y (cdr r)))
(display x) (display y) (newline))
You can generalise this to return multiple values in a list, too.