Singly-linked list/Element insertion: Difference between revisions
m (→{{header|Fortran}}: "elemental" prefix, "value" declared with intent) |
(→{{header|Groovy}}: new solution) |
||
Line 519:
(A C B)
</pre>
=={{header|Groovy}}==
Solution (uses ListNode from [[Singly-Linked List (element)#Groovy]]):
<lang groovy>class NodeList {
private enum Flag { FRONT }
private ListNode head
void insert(value, insertionPoint=Flag.FRONT) {
if (insertionPoint == Flag.FRONT) {
head = new ListNode(payload: value, next: head)
} else {
def node = head
while (node.payload != insertionPoint) {
node = node.next
if (node == null) {
throw new IllegalArgumentException(
"Insertion point ${afterValue} not already contained in list")
}
}
node.next = new ListNode(payload:value, next:node.next)
}
}
String toString() { "${head}" }
}</lang>
Test:
<lang groovy>def list = new NodeList()
list.insert('B')
list.insert('A')
println list
list.insert('C', 'A')
println list</lang>
Output:
<pre>A -> B -> null
A -> C -> B -> null</pre>
=={{header|Haskell}}==
|
Revision as of 09:17, 24 February 2012
You are encouraged to solve this task according to the task description, using any language you may know.
Using the link element defined in Singly-Linked List (element), define a method to insert an element into a singly-linked list following a given element.
Using this method, insert an element C into a list comprised of elements A->B, following element A.
ActionScript
Insertion method: <lang ActionScript>package { public class Node { public var data:Object = null; public var link:Node = null;
public function insert(node:Node):void { node.link = link; link = node; } } }</lang> Usage: <lang ActionScript>import Node;
var A:Node = new Node(1); var B:Node = new Node(2); var C:Node = new Node(3); A.insert(B); A.insert(C);</lang>
Ada
We must create a context clause making the predefined generic procedure Ada.Unchecked_Deallocation visible to this program. <lang ada>with Ada.Unchecked_Deallocation; -- Define the link type procedure Singly_Linked is
type Link; type Link_Access is access Link; type Link is record Data : Integer; Next : Link_Access := null; end record; -- Instantiate the generic deallocator for the link type procedure Free is new Ada.Unchecked_Deallocation(Link, Link_Access);
-- Define the procedure procedure Insert_Append(Anchor : Link_Access; Newbie : Link_Access) is begin if Anchor /= null and Newbie /= null then Newbie.Next := Anchor.Next; Anchor.Next := Newbie; end if; end Insert_Append;
-- Create the link elements A : Link_Access := new Link'(1, null); B : Link_Access := new Link'(2, null); C : Link_Access := new Link'(3, null);
-- Execute the program begin
Insert_Append(A, B); Insert_Append(A, C); Free(A); Free(B); Free(C);
end Singly_Linked;</lang>
ALGOL 68
Linked lists are not built into ALGOL 68 per se, nor any available standard library. However Linked lists are presented in standard text book examples. Or can be manually constructed, eg: <lang algol68>MODE STRINGLIST = STRUCT(STRING value, REF STRINGLIST next);
STRINGLIST list := ("Big",
LOC STRINGLIST := ("fjords", LOC STRINGLIST := ("vex", LOC STRINGLIST := ("quick", LOC STRINGLIST := ("waltz", LOC STRINGLIST := ("nymph",NIL))))));
PROC insert = (REF STRINGLIST list, node)VOID: (
next OF node := next OF list; next OF list := node
);
STRINGLIST very := ("VERY", NIL);
- EXAMPLE OF INSERTION #
insert(next OF next OF list, very );
REF STRINGLIST node := list; WHILE REF STRINGLIST(node) ISNT NIL DO
print((value OF node, space)); node := next OF node
OD; print((newline))</lang>
Output:
Big fjords vex VERY quick waltz nymph
AutoHotkey
<lang AutoHotkey>a = 1 a_next = b b = 2 b_next = 0 c = 3 insert_after("c", "a") Listvars msgbox return
insert_after(new, old) {
local temp temp := %old%_next %old%_next := new %new%_next := temp
}</lang>
BBC BASIC
<lang bbcbasic> DIM node{pNext%, iData%}
DIM a{} = node{}, b{} = node{}, c{} = node{} a.pNext% = b{} a.iData% = 123 b.iData% = 789 c.iData% = 456 PROCinsert(a{}, c{}) END DEF PROCinsert(here{}, new{}) new.pNext% = here.pNext% here.pNext% = new{} ENDPROC
</lang>
C
Define the method:
<lang c>void insert_append (link *anchor, link *newlink) {
newlink->next = anchor->next; anchor->next = newlink;
}</lang>
Note that in a production implementation, one should check anchor and newlink to ensure they're valid values. (I.e., not NULL.)
And now on to the code.
Create our links. <lang c>link *a, *b, *c; a = malloc(sizeof(link)); b = malloc(sizeof(link)); c = malloc(sizeof(link)); a->data = 1; b->data = 2; c->data = 3;</lang>
Prepare our initial list <lang c> insert_append (a, b);</lang>
Insert element c after element a <lang c> insert_append (a, c);</lang>
Remember to free the memory once we're done. <lang c> free (a);
free (b); free (c);</lang>
C++
This uses the generic version of the link node. Of course, normally this would be just some implementation detail inside some list class, not to be used directly by client code.
<lang cpp>template<typename T> void insert_after(link<T>* list_node, link<T>* new_node) {
new_node->next = list_node->next; list_node->next = new_node;
};</lang>
Here's the example code using that method:
The following code creates the links. As numeric values I've just taken the corresponding character values. <lang cpp>link<int>* a = new link<int>('A', new link<int>('B')); link<int>* c = new link<int>('C');</lang>
Now insert c after a: <lang cpp> insert_after(a, c);</lang>
Finally destroy the list: <lang cpp>while (a) {
link<int>* tmp = a; a = a->next; delete tmp;
}</lang>
C#
Creates nodes and inserts them from the data passed. <lang csharp>static void InsertAfter(Link prev, int i) {
prev.next = new Link() { item = i, next = prev.next };
}</lang> <lang csharp>static void Main() {
//Create A(5)->B(7) var A = new Link() { item = 5 }; InsertAfter(A, 7); //Insert C between A and B InsertAfter(A, 15);
}</lang>
Clojure
<lang lisp>(defn insert-after [new old ls]
(cond (empty? ls) ls (= (first ls) old) (cons old (cons new (rest ls))) :else (cons (first ls) (insert-after new old (rest ls)))))</lang>
And the test: <lang lisp>user=> (insert-after 'c 'a '(a b)) (a c b)</lang>
Common Lisp
For many list manipulations in Common Lisp, there are both destructive and non-destructive versions. insert-after
is non-destructive, copying the structure of list up to and including the occurrence of the old-element, and sharing the list structure afterward. ninsert-after
may modify the structure of the input list.
<lang lisp>(defun insert-after (new-element old-element list &key (test 'eql))
"Return a list like list, but with new-element appearing after the
first occurence of old-element. If old-element does not appear in list, then a list returning just new-element is returned."
(if (endp list) (list new-element) (do ((head (list (first list)) (cons (first tail) head)) (tail (rest list) (rest tail))) ((or (endp tail) (funcall test old-element (first head))) (nreconc head (cons new-element tail))))))
(defun ninsert-after (new-element old-element list &key (test 'eql))
"Like insert-after, but modifies list in place. If list is empty, a
new list containing just new-element is returned."
(if (endp list) (list new-element) (do ((prev list next) (next (cdr list) (cdr next))) ((or (null next) (funcall test old-element (car prev))) (rplacd prev (cons new-element next)) list))))</lang>
A simpler implementation that traverses the list a bit more can also be written. This takes advantage of the fact that member returns the tail of the list beginning with the first occurrence of an item, and that ldiff copies as much of its list argument as necessary.
<lang lisp>(defun simple-insert-after (new-element old-element list &key (test 'eql))
(let ((tail (rest (member old-element list :test test)))) (nconc (ldiff list tail) (cons new-element tail))))</lang>
Lastly, here is a recursive version. Case 3 could be optimized by only doing the rplacd operation when the recursive call returns a tail whose first cell is now different compared to that of the previous tail. (I.e. the recursive call has immediately hit case 1 or 2 which allocate new structure.)
<lang lisp>(defun insert-after (list new existing &key (test #'eql)) "Insert item new into list, before existing, or at the end if existing is not present. The default comparison test function is EQL. This function destroys the original list and returns the new list."
(cond ;; case 1: list is empty: just return list of new ((endp list) (list new)) ;; case 2: existing element is first element of list ((funcall test (car list) existing) `(,(car list) ,new ,@(cdr list))) ;; case 3: recurse: insert the element into the rest of the list, ;; and make that list the new rest. (t (rplacd list (insert-before (cdr list) new existing :test test)) list)))</lang>
D
This is similar to C++ example.
<lang D>Node!(T) insert_element_after(T)(Node!(T) dst, Node!(T) src) {
src.next = dst.next; dst.next = src; return dst;
}</lang>
Sample usage (tango based): <lang D>void main() {
alias Node!(char[]) NodeStr; auto a = new NodeStr("alpha", new NodeStr("bravo")); auto c = new NodeStr("charlie");
insert_element_after(a, c);
// print out while (a) { Stdout(a.data) (" -> "); a = a.next; } // gc will handle allocated memory
}</lang>
Delphi
A simple insertion into a one way list. I use a generic pointer for the data that way it can point to any structure, individual variable or whatever. NOTE: For original versions of Turbo Pascal, substitute the MemAvail Function for the Try Except block as this does not exist in this version of the pascal language. Also, Turbo Pascal doesn't have C++-style comments, therefore those have to be replaced with Pascal style comments, i.e. { ... } or (* ... *).
<lang delphi>// Using the same type defs from the one way list example.
Type
// The pointer to the list structure pOneWayList = ^OneWayList;
// The list structure OneWayList = record pData : pointer ; Next : pOneWayList ; end;
// I will illustrate a simple function that will return a pointer to the // new node or it will return NIL. In this example I will always insert // right, to keep the code clear. Since I am using a function all operations // for the new node will be conducted on the functions result. This seems // somewhat counter intuitive, but it is the simplest way to accomplish this.
Function InsertNode(VAR CurrentNode:pOneWayList): pOneWayList begin
// I try not to introduce different parts of the language, and keep each // example to just the code required. in this case it is important to use // a try/except block. In any OS that is multi-threaded and has many apps // running at the same time, you cannot rely on a call to check memory available // and then attempting to allocate. In the time between the two, another // program may have grabbed the memory you were trying to get.
Try // Try to allocate enough memory for a variable the size of OneWayList GetMem(Result,SizeOf(OneWayList)); Except On EOutOfMemoryError do begin Result := NIL exit; end; end;
// Initialize the variable. Result.Next := NIL ; Reuslt.pdata := NIL ;
// Ok now we will insert to the right.
// Is the Next pointer of CurrentNode Nil? If it is we are just tacking // on to the end of the list.
if CurrentNode.Next = NIL then CurrentNode.Next := Result else // We are inserting into the middle of this list Begin Result.Next := CurrentNode.Next ; CurrentNode.Next := result ; end;
end;</lang>
E
<lang e>def insertAfter(head :LinkedList ? (!head.null()),
new :LinkedList ? (new.next().null())) { new.setNext(head.next()) head.setNext(new)
}
def a := makeLink(1, empty) def b := makeLink(2, empty) def c := makeLink(3, empty)
insertAfter(a, b) insertAfter(a, c)
var x := a while (!x.null()) {
println(x.value()) x := x.next()
}</lang>
Factor
<lang factor>: list-append ( previous new -- )
[ swap next>> >>next drop ] [ >>next drop ] 2bi ;
SYMBOLS: A B C ;
A <linked-list> [ C <linked-list> list-append ] keep [ B <linked-list> list-append ] keep .</lang> Output:
T{ linked-list { data A } { next T{ linked-list { data B } { next T{ linked-list { data C } } } } } }
Fantom
Extending Node class from Singly-Linked_List_(element):
<lang fantom> class Node {
const Int value Node? successor // can be null, for end of series
new make (Int value, Node? successor := null) { this.value = value this.successor = successor }
// insert method for this problem public Void insert (Node newNode) { newNode.successor = this.successor this.successor = newNode }
}
// simple class to test putting 'c' between 'a' and 'b' class Main {
public static Void main () { c := Node (2) b := Node (3) a := Node (1, b) a.insert (c)
echo (a.value) echo (a.successor.value) echo (a.successor.successor.value) }
} </lang>
Output:
1 2 3
Forth
Using the linked list concept described in the Singly-Linked_List_(element) topic: <lang forth>\ Create the list and some list elements create A 0 , char A , create B 0 , char B , create C 0 , char C ,</lang>
Now insert b after a and c after b, giving a->b->c <lang forth>B A chain C B chain</lang>
Here is an abbreviated version of the definition of 'chain' from the other article: <lang forth> : chain ( a b -- ) 2dup @ swap ! ! ;</lang>
Fortran
In ISO Fortran 95 or later: <lang fortran>elemental subroutine addAfter(nodeBefore,value)
type (node), intent(inout) :: nodeBefore real, intent(in) :: value type (node), pointer :: newNode allocate(newNode) newNode%data = value newNode%next => nodeBefore%next nodeBefore%next => newNode
end subroutine addAfter</lang>
Go
<lang go>package main
import "fmt"
type Ele struct {
Data interface{} Next *Ele
}
func (e *Ele) insert(data interface{}) {
if e == nil { panic("attept to modify nil") } e.Next = &Ele{data, e.Next}
}
func (e *Ele) printList() {
if e == nil { fmt.Println(nil) return } fmt.Printf("(%v", e.Data) for { e = e.Next if e == nil { fmt.Println(")") return } fmt.Print(" ", e.Data) }
}
func main() {
h := &Ele{"A", &Ele{"B", nil}} h.printList() h.insert("C") h.printList()
}</lang> Output:
(A B) (A C B)
Groovy
Solution (uses ListNode from Singly-Linked List (element)#Groovy): <lang groovy>class NodeList {
private enum Flag { FRONT } private ListNode head void insert(value, insertionPoint=Flag.FRONT) { if (insertionPoint == Flag.FRONT) { head = new ListNode(payload: value, next: head) } else { def node = head while (node.payload != insertionPoint) { node = node.next if (node == null) { throw new IllegalArgumentException( "Insertion point ${afterValue} not already contained in list") } } node.next = new ListNode(payload:value, next:node.next) } } String toString() { "${head}" }
}</lang>
Test: <lang groovy>def list = new NodeList() list.insert('B') list.insert('A') println list
list.insert('C', 'A') println list</lang>
Output:
A -> B -> null A -> C -> B -> null
Haskell
This kind of list manipulation is unidiomatic Haskell. But you can try the following: <lang haskell>insertAfter a b (c:cs) | a==c = a : b : cs
| otherwise = c : insertAfter a b cs
insertAfter _ _ [] = error "Can't insert"</lang>
Icon and Unicon
The Icon solution works for both Icon and Unicon, but Unicon permits a class-based solution.
Icon
<lang Icon> record Node (value, successor)
procedure insert_node (node, newNode)
newNode.successor := node.successor node.successor := newNode
end </lang>
Unicon
<lang Unicon> class Node (value, successor)
method insert (node) node.successor := self.successor self.successor := node end
initially (value, successor) self.value := value self.successor := successor
end </lang>
J
<lang J>list=: 1 65,:_ 66 A=:0 NB. reference into list B=:1 NB. reference into list insertAfter=: monad define
'localListName localListNode localNewValue'=. y localListValue=: ".localListName localOldLinkRef=: <localListNode,0 localNewLinkRef=: #localListValue localNewNode=: (localOldLinkRef { localListValue), localNewValue (localListName)=: (localNewLinkRef localOldLinkRef} localListValue), localNewNode
)</lang>
With these definitions:
insertAfter 'list';A;67
updates the list inserting the value for C after the value for A.
That said, note that the underlying mechanism is rather silly, for J. Linked lists are only interesting in J for illustrative purposes, and should not be used in code that anyone cares about. I have supplied a correspondingly verbose implementation.
Java
Extending Singly-Linked_List_(element)#Java <lang Java>void insertNode(Node<T> anchor_node, Node<T> new_node) {
new_node.next = anchor_node.next; anchor_node.next = new_node;
}</lang>
Java allows the use of generics to allow the data type to be determined at compile time. This will only work on reference types, not primitive types like int or float (wrapper classes like Integer and Float are available).
JavaScript
Extending Singly-Linked_List_(element)#JavaScript <lang javascript>LinkedList.prototype.insertAfter = function(searchValue, nodeToInsert) {
if (this._value == searchValue) { nodeToInsert.next(this.next()); this.next(nodeToInsert); } else if (this.next() == null) throw new Error(0, "value '" + searchValue + "' not found in linked list.") else this.next().insertAfter(searchValue, nodeToInsert);
} var list = createLinkedListFromArray(['A','B']); list.insertAfter('A', new LinkedList('C', null));</lang>
Logo
<lang logo>to insert :after :list :value
localmake "tail member :after :list if not empty? :tail [.setbf :tail fput :value bf :tail] output :list
end
show insert 5 [3 5 1 8] 2</lang>
[3 5 2 1 8]
Mathematica
<lang Mathematica>Append[{a, b}, c] ->{a, b, c}</lang>
OCaml
This kind of list manipulation is unidiomatic OCaml. But you can try the following: <lang ocaml>let rec insert_after a b = function
c :: cs when a = c -> a :: b :: cs | c :: cs -> c :: insert_after a b cs | [] -> raise Not_found</lang>
Pascal
Note: This code uses only Standard Pascal features. For code using features only available in modern Pascal versions, see above under "[Delphi / Object Pascal / >>Turbo Pascal<<]"
Since Standard Pascal doesn't know a generic pointer type, and also no generic types, one has to settle for a specific data type for the linked list. Since the task mentions node names "A", "B", "C", here a char is chosen. Of course any data type (including pointers to a specific data type) could have been used here.
<lang pascal>type
pCharNode = ^CharNode; CharNode = record data: char; next: pCharNode; end;
(* This procedure inserts a node (newnode) directly after another node which is assumed to already be in a list.
It does not allocate a new node, but takes an already allocated node, thus allowing to use it (together with a procedure to remove a node from a list) for splicing a node from one list to another. *)
procedure InsertAfter(listnode, newnode: pCharNode); begin
newnode^.next := listnode^.next; listnode^.next := newnode;
end;</lang> Usage example: <lang pascal>var
A, B: pCharNode;
begin
(* build the two-component list A->C manually *) new(A); A^.data := 'A'; new(A^.next); A^.next^.data := 'C'; A^.next^.next := nil;
(* create the node to be inserted. The initialization of B^.next isn't strictly necessary (it gets overwritten anyway), but it's good style not to leave any values undefined. *) new(B); node^.data := 'B'; node^.next := nil;
(* call the above procedure to insert node B after node A *) InsertAfter(A, B);
(* delete the list *) while A <> nil do begin B := A; A := A^.next; dispose(B); end
end.</lang>
Perl
If you don't really need the constant-time insertion property of singly linked lists, just use an array. You can traverse and splice it any way. <lang perl>my @l = ($A, $B); push @l, $C, splice @l, 1;</lang> However, if you really need a linked list, or all you got is an algorithm in a foreign language, you can use references to accomplish the translation. <lang perl>sub insert_after {
# first argument: node to insert after # second argument: node to insert $_[1]{next} = $_[0]{next}; $_[0]{next} = $_[1];
}
my %B = (
data => 3, next => undef, # not a circular list
); my %A = (
data => 1, next => \%B,
); my %C = (
data => 2,
); insert_after \%A, \%C;</lang> Note that you don't have to name your new nodes. The following works just as well: <lang perl> insert_after \%A, { data => 2 };</lang> Note the curly braces instead of round parentheses.
It is straightforward to extend the function to take an arbitrary number of list nodes to insert: <lang perl>sub insert_after {
my $node = $_[0]; my $next = $node->{next}; shift; while (defined $_[0]) { $node->{next} = $_[0]; $node = $node->{next}; shift; } $node->{next} = $next;
}</lang> With this, it's rather easy to build a list: <lang perl>my %list = ( data => 'A' ); insert_after \%list, { data => 'B' }, { data => 'C' };</lang> List handling is simplified if the variables themselves contain references. For example: <lang perl>my $list2;
- create a new list ('A'. 'B', 'C') and store it in $list2
insert_after $list2 = { data => 'A' }, { data => 'B' }, { data => 'C' };
- append two new nodes ('D', 'E') after the first element
insert_after $list2, { data => 'A2' }, { data => 'A3' };
- append new nodes ('A2a', 'A2b') after the second element (which now is 'A2')
insert_after $list2->{next}, { data => 'A2a' }, { data => 'A2b' };</lang>
Perl 6
<lang perl6>my $letters = 'A' => 'C' => Mu;
sub insert-after($list, $after, $new) {
loop (my $l = $list; $l; $l = $l.value) { if $l.key eqv $after { $l.value = $new => $l.value; return; } } die "Element $after not found";
}
$letters.&insert-after('A', 'B');</lang>
PicoLisp
Destructive operation <lang PicoLisp>(de insertAfter (Item Lst New)
(when (member Item Lst) (con @ (cons New (cdr @))) ) Lst )</lang>
Non-destructive operation <lang PicoLisp>(de insertAfter (Item Lst New)
(if (index Item Lst) (conc (cut @ 'Lst) (cons New Lst)) Lst ) )</lang>
Output in both cases:
: (insertAfter 'A '(A B) 'C) -> (A C B) : (insertAfter 'A '(X Y Z A B D E) 'C) -> (X Y Z A C B D E)
Pop11
In Pop11 one normally uses built-in lists:
<lang pop11>define insert_into_list(anchor, x);
cons(x, back(anchor)) -> back(anchor);
enddefine;
- Build inital list
lvars l1 = cons("a", []); insert_into_list(l1, "b");
- insert c
insert_into_list(l1, "c");</lang>
If one wants one can use user-defined list node (for convenience we repeat definition of list node):
<lang pop11>uses objectclass; define :class ListNode;
slot value = []; slot next = [];
enddefine;
define insert_into_List(anchor, x);
consListNode(x, next(anchor)) -> next(anchor);
enddefine;
- Build inital list
lvars l2 = consListNode("a", []); insert_into_List(l2, "b");
- insert c
insert_into_List(l2, "c");</lang>
Note that user-defined case differs from built-in case only because of names.
PureBasic
<lang PureBasic>Procedure insertAfter(Value, *node.MyData = #Null)
Protected *newNode.MyData = AllocateMemory(SizeOf(MyData)) If *newNode If *node *newNode\next = *node\next *node\next = *newNode EndIf *newNode\Value = Value EndIf ProcedureReturn *newNode ;return pointer to newnode
EndProcedure
Define *SL_List.MyData, a = 1, b = 2, c = 3
- SL_List = insertAfter(a) ;start the list
insertAfter(b, *SL_List) ;insert after head of list insertAfter(c, *SL_List) ;insert after head of list and before tail</lang>
Python
<lang python>def chain_insert(lst, at, item):
while lst is not None: if lst[0] == at: lst[1] = [item, lst[1]] return else: lst = lst[1] raise ValueError(str(at) + " not found")
chain = ['A', ['B', None]] chain_insert(chain, 'A', 'C') print chain</lang> Output: <lang python>['A', ['C', ['B', None]]]</lang>
Ruby
<lang ruby>class ListNode
def insert_after(search_value, new_value) if search_value == value self.succ = self.class.new(new_value, succ) elsif self.succ.nil? raise StandardError, "value #{search_value} not found in list" else self.succ.insert_after(search_value, new_value) end end
end
list = ListNode.new(:a, ListNode.new(:b)) list.insert_after(:a, :c)</lang>
Scala
Placing the method in a companion object (like a static method in Java) <lang scala>object Node {
def insert(a: Node, c: Node) = { c.next = a.next a.next = c }
} </lang>
Scheme
Non-mutating: <lang scheme>(define (insert-after a b lst)
(if (null? lst) lst ; This should be an error, but we will just return the list untouched (let ((c (car lst)) (cs (cdr lst))) (if (equal? a c) (cons a (cons b cs)) (cons c (insert-after a b cs))))))</lang>
Mutating: <lang scheme>(define (insert-after! a b lst)
(let ((pos (member a lst))) (if pos (set-cdr! pos (cons b (cdr pos))))))</lang>
Tcl
<lang tcl># Assume rest of definition is already present oo::define List method insertAfter element {
$element attach $next set next $element
}
set A [List new "A" [List new "B"]] $A insertAfter [List new "C"]</lang>