Largest int from concatenated ints: Difference between revisions
→{{header|Fortran}}: Ah yes, arrays... |
|||
Line 414: | Line 414: | ||
So, the plan is to regard the numbers as being text sequences aligned to the left, containing only digit characters of course - except for the fact that CHARACTER variables often end up having trailing spaces. F2003 formalised a scheme whereby such variables can be "cut-to-fit" as execution proceeds but with earlier Fortrans the standard method is to pay attention to the number of characters in use. F90 introduced a function LEN_TRIM(text) to return the index of the last non-blank character in a text so the only problem now is to decide on how long might the largest number be (and by representing numbers as text strings, there is no difficulty with the limits of INTEGER*2 or INTEGER*4 ''etc.''), and what will be the maximum number of numbers. By devising a subroutine to do the work, these issues can be handled by the caller that is providing the data. The subroutine however intends to sort the collection of texts. This could be done by damaging its parameter which might be regarded as impolite or even unwanted so instead the sort is effected via an array XLAT and juggling its values. This has the advantage that the possibly large elements of the text array are not being moved about, but means that the subroutine must be able to have an XLAT array that is "large enough". F90 standardised the ability for a routine to declare such an array at run-time; previously, arrays within a subroutine (or indeed anywhere) had to have a size fixed at compilation time. In the past this might have been handled by the caller supplying such an array as an additional parameter. |
So, the plan is to regard the numbers as being text sequences aligned to the left, containing only digit characters of course - except for the fact that CHARACTER variables often end up having trailing spaces. F2003 formalised a scheme whereby such variables can be "cut-to-fit" as execution proceeds but with earlier Fortrans the standard method is to pay attention to the number of characters in use. F90 introduced a function LEN_TRIM(text) to return the index of the last non-blank character in a text so the only problem now is to decide on how long might the largest number be (and by representing numbers as text strings, there is no difficulty with the limits of INTEGER*2 or INTEGER*4 ''etc.''), and what will be the maximum number of numbers. By devising a subroutine to do the work, these issues can be handled by the caller that is providing the data. The subroutine however intends to sort the collection of texts. This could be done by damaging its parameter which might be regarded as impolite or even unwanted so instead the sort is effected via an array XLAT and juggling its values. This has the advantage that the possibly large elements of the text array are not being moved about, but means that the subroutine must be able to have an XLAT array that is "large enough". F90 standardised the ability for a routine to declare such an array at run-time; previously, arrays within a subroutine (or indeed anywhere) had to have a size fixed at compilation time. In the past this might have been handled by the caller supplying such an array as an additional parameter. |
||
Passing arrays as parameters can be tricky, especially for multi-dimensional arrays. This uses the old style whereby the size is left unstated via the * in <code>TEXT(*)</code>, though one could use <code>TEXT(N)</code> instead - but at the risk that the actual value of N is wrong and array index checking might be confused thereby. Still earlier one would simply place some integer value there, any valid integer, as in <code>TEXT(666)</code>, and not worry about bound checking at all because old-style compilers did not produce checking code even if it was wanted. F90 standardised the MODULE protocol, within which the size is specified as <code>TEXT(:)</code> whereby secret additional parameters are supplied that contain the actual bound information and bound checking will be correct, possibly not so if the <code>TEXT(N)</code> form is used instead and N is wrong. This extra overhead in every use is possibly better than undetected errors in some uses... |
|||
The sorting of the text array was to be by the notorious BubbleSort, taking advantage of the fact that each pass delivers the maximum value of the unsorted portion to its final position: the output could thereby be produced as the sort worked. Rather than mess about with early termination (no element being swapped) or attention to the bounds within which swapping took place, attention concentrated upon the comparison. Because of the left-alignment of the texts, a simple comparison seemed sufficient until I thought of unequal text lengths and then the following example. Suppose there are two numbers, 5, and one of 54, 55, or 56 as the other. Via normal comparisons, the 5 would always be first (because short texts are considered expanded with trailing spaces when compared against longer texts, and a space precedes every digit) however the biggest ordering is 5 54 for the first case but 56 5 for the last. This possibility is not exemplified in the specified examples. So, a more complex comparison is required. One could of course write a suitable function and consider the issue there but instead the comparison forms the compound text in the same manner as the result will be, in the two ways AB and BA, and looks to see which yields the bigger sequence. This need only be done for unequal length text pairs. |
The sorting of the text array was to be by the notorious BubbleSort, taking advantage of the fact that each pass delivers the maximum value of the unsorted portion to its final position: the output could thereby be produced as the sort worked. Rather than mess about with early termination (no element being swapped) or attention to the bounds within which swapping took place, attention concentrated upon the comparison. Because of the left-alignment of the texts, a simple comparison seemed sufficient until I thought of unequal text lengths and then the following example. Suppose there are two numbers, 5, and one of 54, 55, or 56 as the other. Via normal comparisons, the 5 would always be first (because short texts are considered expanded with trailing spaces when compared against longer texts, and a space precedes every digit) however the biggest ordering is 5 54 for the first case but 56 5 for the last. This possibility is not exemplified in the specified examples. So, a more complex comparison is required. One could of course write a suitable function and consider the issue there but instead the comparison forms the compound text in the same manner as the result will be, in the two ways AB and BA, and looks to see which yields the bigger sequence. This need only be done for unequal length text pairs. |
||
Line 451: | Line 453: | ||
ELSE !But if not, construct the actual candidate texts for comparison. |
ELSE !But if not, construct the actual candidate texts for comparison. |
||
IF (TEXT(TI)(1:LI)//TEXT(TJ)(1:LJ) !These two will be the same length. |
IF (TEXT(TI)(1:LI)//TEXT(TJ)(1:LJ) !These two will be the same length. |
||
1 .LT.TEXT(TJ)(1:LJ)//TEXT(TI)(1:LI)) !Just as above. |
1 .LT.TEXT(TJ)(1:LJ)//TEXT(TI)(1:LI)) !Just as above. |
||
2 CALL SWAP(XLAT(I),XLAT(J)) !J shall now follow I. |
2 CALL SWAP(XLAT(I),XLAT(J)) !J shall now follow I. |
||
END IF !So much for that comparison. |
END IF !So much for that comparison. |
||
Line 461: | Line 463: | ||
PROGRAM POKE |
PROGRAM POKE |
||
CHARACTER*4 T1(10) |
CHARACTER*4 T1(10) !Prepare some example arrays. |
||
CHARACTER*4 T2(4) |
CHARACTER*4 T2(4) !To hold the specified examples. |
||
INTEGER MSG |
INTEGER MSG |
||
COMMON /IODEV/ MSG |
COMMON /IODEV/ MSG |
||
DATA T1(1:8)/"1","34","3","98","9","76","45","4"/ |
DATA T1(1:8)/"1","34","3","98","9","76","45","4"/ |
||
DATA T2/"54","546","548","60"/ |
DATA T2/"54","546","548","60"/ |
||
MSG = 6 |
MSG = 6 !Standard output. |
||
WRITE (MSG,1) |
WRITE (MSG,1) |
||
1 FORMAT ("Takes a list of integers and concatenates them so as ", |
1 FORMAT ("Takes a list of integers and concatenates them so as ", |
||
Line 489: | Line 491: | ||
END DO !Thus, the numbers are aligned left in the text field. |
END DO !Thus, the numbers are aligned left in the text field. |
||
CALL BIGUP(T1,10) |
CALL BIGUP(T1,10) |
||
END</lang> |
END </lang> |
||
Output: the Fortran compiler ignores spaces when reading fortran source, so hard-core fortranners should have no difficulty doing |
Output: the Fortran compiler ignores spaces when reading fortran source, so, hard-core fortranners should have no difficulty doing likewise for the output... |
||
<pre> |
<pre> |
||
Takes a list of integers and concatenates them so as to produce the biggest possible number. |
Takes a list of integers and concatenates them so as to produce the biggest possible number. |
Revision as of 05:09, 25 May 2016
![Task](http://static.miraheze.org/rosettacodewiki/thumb/b/ba/Rcode-button-task-crushed.png/64px-Rcode-button-task-crushed.png)
You are encouraged to solve this task according to the task description, using any language you may know.
Given a set of positive integers, the task is to write a function to order the integers in such a way that the concatenation of the numbers forms the largest possible integer and return this integer.
Use the following two sets of integers as tests and show your program output here.
- {1, 34, 3, 98, 9, 76, 45, 4}
- {54, 546, 548, 60}
- Possible algorithms
- A solution could be found by trying all combinations and return the best.
- Another way to solve this is to note that in the best arrangement, for any two adjacent original integers X and Y, the concatenation X followed by Y will be numerically greater than or equal to the concatenation Y followed by X.
- Yet another way to solve this is to pad ints to the same size by repeating the digits then sort using these repeated ints as a sort key.
- Cf
- Algorithms: What is the most efficient way to arrange the given numbers to form the biggest number?.
- Constructing the largest number possible by rearranging a list
Ada
The algorithmic idea is to apply a twisted comparison function:
<lang Ada>function Order(Left, Right: Natural) return Boolean is
( (Img(Left) & Img(Right)) > (Img(Right) & Img(Left)) );</lang>
This function converts the parameters Left and Right to strings and returns True if (Left before Right) exceeds (Right before Left). It needs Ada 2012 -- the code for older versions of Ada would be more verbose.
The rest is straightforward: Run your favourite sorting subprogram that allows to use the function "Order" instead of standard comparison operators ("<" or ">" or so) and print the results:
<lang Ada>with Ada.Text_IO, Ada.Containers.Generic_Array_Sort;
procedure Largest_Int_From_List is
function Img(N: Natural) return String is S: String := Integer'Image(N); begin return S(S'First+1 .. S'Last); -- First character is ' ' end Img; function Order(Left, Right: Natural) return Boolean is ( (Img(Left) & Img(Right)) > (Img(Right) & Img(Left)) ); type Arr_T is array(Positive range <>) of Natural; procedure Sort is new Ada.Containers.Generic_Array_Sort (Positive, Natural, Arr_T, Order); procedure Print_Sorted(A: Arr_T) is B: Arr_T := A; begin Sort(B); for Number of B loop
Ada.Text_IO.Put(Img(Number));
end loop; Ada.Text_IO.New_Line; end Print_Sorted;
begin
Print_Sorted((1, 34, 3, 98, 9, 76, 45, 4)); Print_Sorted((54, 546, 548, 60));
end Largest_Int_From_List;</lang>
AutoHotkey
<lang AutoHotkey>LargestConcatenatedInts(var){ StringReplace, var, A_LoopField,%A_Space%,, all Sort, var, D`, fConcSort StringReplace, var, var, `,,, all return var }
ConcSort(a, b){ m := a . b , n := b . a
return m < n ? 1 : m > n ? -1 : 0
}</lang> Examples:<lang AutoHotkey>d = ( 1, 34, 3, 98, 9, 76, 45, 4 54, 546, 548, 60 4 , 45, 54, 5 ) loop, parse, d, `n MsgBox % LargestConcatenatedInts(A_LoopField)</lang>
- Output:
998764543431 6054854654 554454
AWK
<lang awk> function cmp(i1, v1, i2, v2, u1, u2) { u1 = v1""v2; u2 = v2""v1;
return (u2 - u1)
} function largest_int_from_concatenated_ints(X) {
PROCINFO["sorted_in"]="cmp";
u=""; for (i in X) u=u""X[i]; return u }
BEGIN { split("1 34 3 98 9 76 45 4",X); print largest_int_from_concatenated_ints(X)
split("54 546 548 60",X); print largest_int_from_concatenated_ints(X) } </lang>
- Output:
998764543431 6054854654
BBC BASIC
<lang bbcbasic> DIM Nums%(10)
Nums%()=1,34,3,98,9,76,45,4 PRINT FNlargestint(8) Nums%()=54,546,548,60 PRINT FNlargestint(4) END DEF FNlargestint(len%) LOCAL i%,l$,a$,b$,sorted% REPEAT sorted%=TRUE FOR i%=0 TO len%-2 a$=STR$Nums%(i%) b$=STR$Nums%(i%+1) IF a$+b$<b$+a$ SWAP Nums%(i%),Nums%(i%+1):sorted%=FALSE NEXT UNTIL sorted% FOR i%=0 TO len%-1 l$+=STR$Nums%(i%) NEXT =l$</lang>
- Output:
998764543431 6054854654
Bracmat
<lang bracmat>( ( maxnum
= A Z F C . !arg:# | !arg : %@?F ? ( #%@?C & ( str$(!F !C)+-1*str$(!C !F):~<0 | !C:?F ) & ~ ) ? | !arg:?A !F ?Z&!F maxnum$(!A !Z) )
& out$(str$(maxnum$(1 34 3 98 9 76 45 4))) & out$(str$(maxnum$(54 546 548 60))) );</lang>
- Output:
998764543431 6054854654
C
<lang C>#include <stdio.h>
- include <stdlib.h>
- include <string.h>
int catcmp(const void *a, const void *b) { char ab[32], ba[32]; sprintf(ab, "%d%d", *(int*)a, *(int*)b); sprintf(ba, "%d%d", *(int*)b, *(int*)a); return strcmp(ba, ab); }
void maxcat(int *a, int len) { int i; qsort(a, len, sizeof(int), catcmp); for (i = 0; i < len; i++) printf("%d", a[i]); putchar('\n'); }
int main(void) { int x[] = {1, 34, 3, 98, 9, 76, 45, 4}; int y[] = {54, 546, 548, 60};
maxcat(x, sizeof(x)/sizeof(x[0])); maxcat(y, sizeof(y)/sizeof(y[0]));
return 0; }</lang>
- Output:
998764543431 6054854654
C++
<lang cpp>#include <iostream>
- include <sstream>
- include <algorithm>
- include <vector>
- include <string>
std::string findLargestConcat ( std::vector< int > & mynumbers ) {
std::vector<std::string> concatnumbers ; std::sort ( mynumbers.begin( ) , mynumbers.end( ) ) ; do { std::ostringstream numberstream ; for ( int i : mynumbers )
numberstream << i ;
concatnumbers.push_back( numberstream.str( ) ) ; } while ( std::next_permutation( mynumbers.begin( ) ,
mynumbers.end( ) )) ;
return *( std::max_element( concatnumbers.begin( ) ,
concatnumbers.end( ) ) ) ; }
int main( ) {
std::vector<int> mynumbers = { 98, 76 , 45 , 34, 9 , 4 , 3 , 1 } ; std::vector<int> othernumbers = { 54 , 546 , 548 , 60 } ; std::cout << "The largest concatenated int is " << findLargestConcat( mynumbers ) << " !\n" ; std::cout << "And here it is " << findLargestConcat( othernumbers ) << " !\n" ; return 0 ;
}</lang>
- Output:
The largest concatenated int is 998764543431 ! And here it is 6054854654 !
C#
<lang csharp>using System; using System.Collections.Generic; using System.Linq;
class Program {
static void Main(string[] args) { var source1 = new int[] { 1, 34, 3, 98, 9, 76, 45, 4 }; var source2 = new int[] { 54, 546, 548, 60 };
var largest1 = LargestPossibleSequence(source1); var largest2 = LargestPossibleSequence(source2);
Console.WriteLine("The largest possible integer from set 1 is: {0}", largest1); Console.WriteLine("The largest possible integer from set 2 is: {0}", largest2); }
static long LargestPossibleSequence(int[] ints) { return long.Parse(string.Join("", ints.OrderBy(i => i, new IntConcatenationComparer()).Reverse())); }
}
class IntConcatenationComparer : IComparer<int> {
public int Compare(int x, int y) { var xy = int.Parse(x.ToString() + y.ToString()); var yx = int.Parse(y.ToString() + x.ToString());
return xy - yx; }
} </lang>
- Output:
The largest possible integer from set 1 is: 998764543431 The largest possible integer from set 2 is: 6054854654
Ceylon
<lang ceylon>shared void run2() {
function intConcatenationComparer(Integer x, Integer y) { assert(exists xy = parseInteger(x.string + y.string), exists yx = parseInteger(y.string + x.string)); return yx <=> xy; }
function biggestConcatenation(Integer* ints) => "".join(ints.sort(intConcatenationComparer));
value test1 = {1, 34, 3, 98, 9, 76, 45, 4}; value test2 = {54, 546, 548, 60};
print("``biggestConcatenation(*test1)`` and ``biggestConcatenation(*test2)``"); }</lang>
Clojure
<lang Clojure>(defn maxcat [coll]
(read-string (apply str (sort (fn [x y] (apply compare (map read-string [(str y x) (str x y)]))) coll))))
(prn (map maxcat [[1 34 3 98 9 76 45 4] [54 546 548 60]]))</lang>
- Output:
(998764543431 6054854654)
Common Lisp
<lang lisp>
- Sort criteria is by most significant digit with least digits used as a tie
- breaker
(defun largest-msd-with-less-digits (x y)
(flet ((first-digit (x) (digit-char-p (aref x 0)))) (cond ((> (first-digit x) (first-digit y)) t) ((> (first-digit y) (first-digit x)) nil) ((and (= (first-digit x) (first-digit y)) (> (length x) (length y))) nil) (t t))))
(loop
:for input :in '((54 546 548 60) (1 34 3 98 9 76 45 4)) :do (format t "~{~A~}~%" (sort (mapcar #'write-to-string input) #'largest-msd-with-less-digits)))
</lang>
- Output:
6054548546 998764453341
D
The three algorithms. Uses the second module from the Permutations Task. <lang d>import std.stdio, std.algorithm, std.conv, std.array, permutations2;
auto maxCat1(in int[] arr) pure @safe {
return arr.to!(string[]).permutations.map!join.reduce!max;
}
auto maxCat2(in int[] arr) pure nothrow @safe {
return arr.to!(string[]).sort!q{b ~ a < a ~ b}.join;
}
auto maxCat3(in int[] arr) /*pure nothrow @safe*/ {
immutable maxL = arr.reduce!max.text.length; return arr.to!(string[]) .schwartzSort!(s => s.replicate(maxL/s.length + 1), "a > b") .join;
}
void main() {
const lists = [[1, 34, 3, 98, 9, 76, 45, 4], [54, 546, 548, 60]]; [&maxCat1, &maxCat2, &maxCat3].map!(cat => lists.map!cat).writeln;
}</lang>
- Output:
[["998764543431", "6054854654"], ["998764543431", "6054854654"], ["998764543431", "6054854654"]]
Elixir
<lang elixir>defmodule RC do
def largest_int(list) do sorted = Enum.sort(list, fn x,y -> "#{x}#{y}" >= "#{y}#{x}" end) Enum.join(sorted) end
end
IO.inspect RC.largest_int [1, 34, 3, 98, 9, 76, 45, 4] IO.inspect RC.largest_int [54, 546, 548, 60]</lang>
- Output:
"998764543431" "6054854654"
Erlang
<lang Erlang> -module( largest_int_from_concatenated ).
-export( [ints/1, task/0] ).
ints( Ints ) -> Int_strings = [erlang:integer_to_list(X) || X <- Ints], Pad_ints = [{X ++ X, X} || X <- Int_strings], erlang:list_to_integer( lists:append([Int || {_Pad, Int} <- lists:reverse(lists:sort(Pad_ints))]) ).
task() -> [io:fwrite("Largest ~p from ~p~n", [ints(X), X]) || X <- [[1, 34, 3, 98, 9, 76, 45, 4], [54, 546, 548, 60]]]. </lang>
- Output:
8> largest_int_from_concatenated:task(). Largest 998764543431 from [1,34,3,98,9,76,45,4] Largest 6054854654 from [54,546,548,60]
Fortran
There is often a potential ambiguity when reading numbers. While three definitely names the Platonic number notion, 3 might instead be regarded as being a text that happens to have the glyph of a number but is not a number. This sort of discussion arises when a spreadsheet has read in a text file and behold! numbers are on the display and they look just like what is displayed when numbers are being shown, but, they are not numbers, they are only drawn that way. Within the spreadsheet they are parts of some text, and the notion that takes over is one of a "blunt, heavy object", not alas close to hand.
So, the plan is to regard the numbers as being text sequences aligned to the left, containing only digit characters of course - except for the fact that CHARACTER variables often end up having trailing spaces. F2003 formalised a scheme whereby such variables can be "cut-to-fit" as execution proceeds but with earlier Fortrans the standard method is to pay attention to the number of characters in use. F90 introduced a function LEN_TRIM(text) to return the index of the last non-blank character in a text so the only problem now is to decide on how long might the largest number be (and by representing numbers as text strings, there is no difficulty with the limits of INTEGER*2 or INTEGER*4 etc.), and what will be the maximum number of numbers. By devising a subroutine to do the work, these issues can be handled by the caller that is providing the data. The subroutine however intends to sort the collection of texts. This could be done by damaging its parameter which might be regarded as impolite or even unwanted so instead the sort is effected via an array XLAT and juggling its values. This has the advantage that the possibly large elements of the text array are not being moved about, but means that the subroutine must be able to have an XLAT array that is "large enough". F90 standardised the ability for a routine to declare such an array at run-time; previously, arrays within a subroutine (or indeed anywhere) had to have a size fixed at compilation time. In the past this might have been handled by the caller supplying such an array as an additional parameter.
Passing arrays as parameters can be tricky, especially for multi-dimensional arrays. This uses the old style whereby the size is left unstated via the * in TEXT(*)
, though one could use TEXT(N)
instead - but at the risk that the actual value of N is wrong and array index checking might be confused thereby. Still earlier one would simply place some integer value there, any valid integer, as in TEXT(666)
, and not worry about bound checking at all because old-style compilers did not produce checking code even if it was wanted. F90 standardised the MODULE protocol, within which the size is specified as TEXT(:)
whereby secret additional parameters are supplied that contain the actual bound information and bound checking will be correct, possibly not so if the TEXT(N)
form is used instead and N is wrong. This extra overhead in every use is possibly better than undetected errors in some uses...
The sorting of the text array was to be by the notorious BubbleSort, taking advantage of the fact that each pass delivers the maximum value of the unsorted portion to its final position: the output could thereby be produced as the sort worked. Rather than mess about with early termination (no element being swapped) or attention to the bounds within which swapping took place, attention concentrated upon the comparison. Because of the left-alignment of the texts, a simple comparison seemed sufficient until I thought of unequal text lengths and then the following example. Suppose there are two numbers, 5, and one of 54, 55, or 56 as the other. Via normal comparisons, the 5 would always be first (because short texts are considered expanded with trailing spaces when compared against longer texts, and a space precedes every digit) however the biggest ordering is 5 54 for the first case but 56 5 for the last. This possibility is not exemplified in the specified examples. So, a more complex comparison is required. One could of course write a suitable function and consider the issue there but instead the comparison forms the compound text in the same manner as the result will be, in the two ways AB and BA, and looks to see which yields the bigger sequence. This need only be done for unequal length text pairs.
The source is F77 style, except for the declaration of XLAT(N), the use of <N> in the FORMAT statements instead of some large constant or similar, and the ability to declare an array via constants as in (/"5","54"/)
rather than mess about declaring arrays and initialising them separately. The I0
format to convert a number (an actual number) into a digit string aligned leftwards in a CHARACTER variable of sufficient size is also a F90 introduction, though the B6700 compiler allowed a code J
instead. This last is to demonstrate usage of actual numbers for those unpersuaded by the argument for ambiguity that allows for texts. If the I0
format code is unavailable then I9
(or some suitable size) could be used, followed by text = ADJUSTL(text)
, except that this became an intrinsic function only in F90, so perhaps you will have to write a simple alignment routine. <lang Fortran> SUBROUTINE SWAP(A,B) !Why can't the compiler supply these!
INTEGER A,B,T T = B B = A A = T END
SUBROUTINE BIGUP(TEXT,N) !Outputs the numbers in TEXT to give the biggest number. CHARACTER*(*) TEXT(*) !The numbers as text, aligned left. INTEGER N !The number of them. INTEGER XLAT(N),L(N) !An index and a set of lengths. INTEGER I,J,M !Assorted steppers. INTEGER TI,TJ !Fingers to a text. INTEGER LI,LJ !Lengths of the fingered texts. INTEGER MSG !I/O unit number. COMMON /IODEV/ MSG !Old style. DO I = 1,N !Step through my supply of texts. XLAT(I) = I !Preparing a finger to them. L(I) = LEN_TRIM(TEXT(I)) !And noting their last non-blank. END DO !On to the next. WRITE (MSG,1) "Supplied",(TEXT(I)(1:L(I)), I = 1,N) !Show the grist. 1 FORMAT (A12,":",<N>(A,",")) !Instead of <N>, 666 might suffice.
Crude bubblesort. No attempt at noting the bounds of swaps made.
DO M = N,1,-1 !Just for fun, go backwards. DO I = 2,M !Start a scan. J = I - 1 !Comparing element I to element I - 1. TI = XLAT(I) !Thus finger the I'th text in XLAT order. TJ = XLAT(J) !And its supposed predecessor. LI = L(TI) !The length of the fingered text. LJ = L(TJ) !All this to save on typing below. IF (LI .EQ. LJ) THEN !If the texts are equal lengths, IF (TEXT(TI).LT.TEXT(TJ)) CALL SWAP(XLAT(I),XLAT(J)) !A simple comparison. ELSE !But if not, construct the actual candidate texts for comparison. IF (TEXT(TI)(1:LI)//TEXT(TJ)(1:LJ) !These two will be the same length. 1 .LT.TEXT(TJ)(1:LJ)//TEXT(TI)(1:LI)) !Just as above. 2 CALL SWAP(XLAT(I),XLAT(J)) !J shall now follow I. END IF !So much for that comparison. END DO !On to the next. END DO !The original plan was to reveal element XLAT(M) as found. WRITE (MSG,2) "Biggest",(TEXT(XLAT(I))(1:L(XLAT(I))),I = N,1,-1) !But, all at once is good too. 2 FORMAT (A12,":",<N>(A," ")) !The space maintains identity. END !That was fun.
PROGRAM POKE CHARACTER*4 T1(10) !Prepare some example arrays. CHARACTER*4 T2(4) !To hold the specified examples. INTEGER MSG COMMON /IODEV/ MSG DATA T1(1:8)/"1","34","3","98","9","76","45","4"/ DATA T2/"54","546","548","60"/ MSG = 6 !Standard output. WRITE (MSG,1) 1 FORMAT ("Takes a list of integers and concatenates them so as ", 1 "to produce the biggest possible number.",/, 2 "The result is shown with spaces between the parts ", 3 "to show provenance. Ignore them otherwise."/) CALL BIGUP(T1,8)
WRITE (MSG,*) CALL BIGUP(T2,4)
WRITE (MSG,*) "These are supplied in lexicographical order..." CALL BIGUP((/"5","54"/),2)
WRITE (MSG,*) "But this is not necessarily the biggest order." CALL BIGUP((/"5","56"/),2)
WRITE (MSG,*) "And for those who count..." DO I = 1,10 WRITE (T1(I),"(I0)") I !This format code produces only the necessary text. END DO !Thus, the numbers are aligned left in the text field. CALL BIGUP(T1,10) END </lang>
Output: the Fortran compiler ignores spaces when reading fortran source, so, hard-core fortranners should have no difficulty doing likewise for the output...
Takes a list of integers and concatenates them so as to produce the biggest possible number. The result is shown with spaces between the parts to show provenance. Ignore them otherwise. Supplied:1,34,3,98,9,76,45,4, Biggest:9 98 76 45 4 34 3 1 Supplied:54,546,548,60, Biggest:60 548 546 54 These are supplied in lexicographical order... Supplied:5,54, Biggest:5 54 But this is not necessarily the biggest order. Supplied:5,56, Biggest:56 5 And for those who count... Supplied:1,2,3,4,5,6,7,8,9,10, Biggest:9 8 7 6 5 4 3 2 1 10
Go
<lang go>// Variation of method 3. Repeat digits to at least the size of the longest, // then sort as strings. package main
import (
"fmt" "math/big" "sort" "strconv" "strings"
)
type c struct {
i int s, rs string
}
type cc []*c
func (c cc) Len() int { return len(c) } func (c cc) Less(i, j int) bool { return c[j].rs < c[i].rs } func (c cc) Swap(i, j int) { c[i], c[j] = c[j], c[i] }
// Function required by task. Takes a list of integers, returns big int. func li(is ...int) *big.Int {
ps := make(cc, len(is)) ss := make([]c, len(is)) ml := 0 for j, i := range is { p := &ss[j] ps[j] = p p.i = i p.s = strconv.Itoa(i) if len(p.s) > ml { ml = len(p.s) } } for _, p := range ps { p.rs = strings.Repeat(p.s, (ml+len(p.s)-1)/len(p.s)) } sort.Sort(ps) s := make([]string, len(ps)) for i, p := range ps { s[i] = p.s } b, _ := new(big.Int).SetString(strings.Join(s, ""), 10) return b
}
func main() {
fmt.Println(li(1, 34, 3, 98, 9, 76, 45, 4)) fmt.Println(li(54, 546, 548, 60))
}</lang>
- Output:
998764543431 6054854654
Go
<lang go>// Variation of method 3. Repeat digits to at least the size of the longest, // then sort as strings. package main
import (
"fmt" "math/big" "sort" "strconv" "strings"
)
type c struct {
i int s, rs string
}
type cc []*c
func (c cc) Len() int { return len(c) } func (c cc) Less(i, j int) bool { return c[j].rs < c[i].rs } func (c cc) Swap(i, j int) { c[i], c[j] = c[j], c[i] }
// Function required by task. Takes a list of integers, returns big int. func li(is ...int) *big.Int {
ps := make(cc, len(is)) ss := make([]c, len(is)) ml := 0 for j, i := range is { p := &ss[j] ps[j] = p p.i = i p.s = strconv.Itoa(i) if len(p.s) > ml { ml = len(p.s) } } for _, p := range ps { p.rs = strings.Repeat(p.s, (ml+len(p.s)-1)/len(p.s)) } sort.Sort(ps) s := make([]string, len(ps)) for i, p := range ps { s[i] = p.s } b, _ := new(big.Int).SetString(strings.Join(s, ""), 10) return b
}
func main() {
fmt.Println(li(1, 34, 3, 98, 9, 76, 45, 4)) fmt.Println(li(54, 546, 548, 60))
}</lang>
- Output:
998764543431 6054854654
Groovy
<lang groovy>def largestInt = { c -> c.sort { v2, v1 -> "$v1$v2" <=> "$v2$v1" }.join() as BigInteger }</lang> Testing: <lang groovy>assert largestInt([1, 34, 3, 98, 9, 76, 45, 4]) == 998764543431 assert largestInt([54, 546, 548, 60]) == 6054854654</lang>
Haskell
Compare repeated string method
<lang Haskell>import Data.List (sortBy) import Data.Ord (comparing)
main = print (map maxcat [[1,34,3,98,9,76,45,4], [54,546,548,60]] :: [Integer])
where sorted xs = let pad x = concat $ replicate (maxLen `div` length x + 1) x maxLen = maximum $ map length xs in sortBy (flip $ comparing pad) xs
maxcat = read . concat . sorted . map show</lang>
- Output:
[998764543431,6054854654]
Since repeating numerical string "1234" is the same as taking all the digits of 1234/9999 after the decimal point, the following does essentially the same as above: <lang haskell>import Data.List (sortBy) import Data.Ord (comparing) import Data.Ratio ((%))
nines = iterate ((+9).(*10)) 9
maxcat = foldl (\a (n,d)->a * (1 + d) + n) 0 .
sortBy (flip $ comparing $ uncurry (%)) . map (\a->(a, head $ dropWhile (<a) nines))
main = mapM_ (print.maxcat) [[1,34,3,98,9,76,45,4], [54,546,548,60]]</lang>
Sort on comparison of concatenated ints method
<lang Haskell>import Data.List (sortBy)
main = print (map maxcat [[1,34,3,98,9,76,45,4], [54,546,548,60]] :: [Integer])
where sorted = sortBy (\a b -> compare (b++a) (a++b)) maxcat = read . concat . sorted . map show</lang>
- Output as above.
Try all permutations method
<lang Haskell>import Data.List (sort, permutations)
main = print (map maxcat [[1,34,3,98,9,76,45,4], [54,546,548,60]] :: [Integer])
where maxcat = read . last . sort . map (concat . map show) . permutations</lang>
- Output as above.
Icon and Unicon
This solution only works in Unicon as it uses a Heap class to do the heavy lifting.
<lang unicon>import Collections # For the Heap (dense priority queue) class
procedure main(a)
write(lici(a))
end
procedure lici(a)
every (result := "") ||:= Heap(a,,cmp).gen() return result
end
procedure cmp(a,b)
return (a||b) > (b||a)
end</lang>
Sample runs:
->lici 1 34 3 98 9 76 45 4 998764543431 ->lici 54 546 548 60 6054854654 ->
J
Solution: <lang j>maxlen=: [: >./ #&> maxnum=: (0 ". ;)@(\: maxlen $&> ])@(8!:0)</lang> Usage: <lang j> maxnum&> 1 34 3 98 9 76 45 4 ; 54 546 548 60 998764543431 6054854654</lang>
Java
This example sets up a comparator to order the numbers using Collections.sort
as described in method #3 (padding and reverse sorting).
It was also necessary to make a join method to meet the output requirements.
<lang java5>import java.util.*;
public class IntConcat {
private static Comparator<Integer> sorter = new Comparator<Integer>(){ @Override public int compare(Integer o1, Integer o2){ String o1s = o1.toString(); String o2s = o2.toString(); if(o1s.length() == o2s.length()){ return o2s.compareTo(o1s); }
int mlen = Math.max(o1s.length(), o2s.length()); while(o1s.length() < mlen * 2) o1s += o1s; while(o2s.length() < mlen * 2) o2s += o2s; return o2s.compareTo(o1s); } }; public static String join(List<?> things){ String output = ""; for(Object obj:things){ output += obj; } return output; } public static void main(String[] args){ List<Integer> ints1 = new ArrayList<Integer>(Arrays.asList(1, 34, 3, 98, 9, 76, 45, 4)); Collections.sort(ints1, sorter); System.out.println(join(ints1)); List<Integer> ints2 = new ArrayList<Integer>(Arrays.asList(54, 546, 548, 60)); Collections.sort(ints2, sorter); System.out.println(join(ints2)); }
}</lang>
<lang java5>import java.util.Comparator; import java.util.stream.Collectors; import java.util.stream.Stream;
public interface IntConcat {
public static Comparator<Integer> SORTER = (o1, o2) -> { String o1s = o1.toString(); String o2s = o2.toString(); if (o1s.length() == o2s.length()) { return o2s.compareTo(o1s); } int mlen = Math.max(o1s.length(), o2s.length()); while (o1s.length() < mlen * 2) { o1s += o1s; } while (o2s.length() < mlen * 2) { o2s += o2s; } return o2s.compareTo(o1s); };
public static void main(String[] args) { Stream<Integer> ints1 = Stream.of( 1, 34, 3, 98, 9, 76, 45, 4 );
System.out.println(ints1 .parallel() .sorted(SORTER) .map(String::valueOf) .collect(Collectors.joining()) );
Stream<Integer> ints2 = Stream.of( 54, 546, 548, 60 );
System.out.println(ints2 .parallel() .sorted(SORTER) .map(String::valueOf) .collect(Collectors.joining()) ); }
}</lang>
- Output:
998764543431 6054854654
JavaScript
ES5
<lang JavaScript> (function () {
'use strict';
// maxCombine :: [Int] -> Int function maxCombine(xs) { return parseInt( xs.sort( function (x, y) { var a = x.toString(), b = y.toString(), ab = parseInt(a + b), ba = parseInt(b + a);
return ab > ba ? -1 : (ab < ba ? 1 : 0); } ) .join(), 10 ); }
return [ [1, 34, 3, 98, 9, 76, 45, 4], [54, 546, 548, 60] ].map(maxCombine);
})();
</lang>
- Output:
[998764543431, 6054854654]
jq
Padding
For jq versions greater than 1.4, it may be necessary to change "sort_by" to "sort". <lang jq>def largest_int:
def pad(n): . + (n - length) * .[length-1:];
map(tostring) | (map(length) | max) as $max | map([., pad($max)]) | sort_by( .[1] ) | map( .[0] ) | reverse | join("") ;
- Examples:
([1, 34, 3, 98, 9, 76, 45, 4],
[54, 546, 548, 60]) | largest_int
</lang>
- Output:
$ /usr/local/bin/jq -n -M -r -f Largest_int_from_concatenated_ints.jq 998764543431 6054854654
Custom Sort
The following uses quicksort/1: <lang jq>def largest_int:
map(tostring) | quicksort( .[0] + .[1] < .[1] + .[0] ) | reverse | join("") ;</lang>
Julia
Perhaps algorithm 3 is more efficient, but algorithm 2 is decent and very easy to implement in Julia. So this solution uses algorithm 2. <lang Julia> function maxconcat{T<:Integer}(a::Array{T,1})
b = map(string, a) b = sort(b, lt=(x,y)->x*y < y*x, rev=true) b = join(b, "") try b = parseint(b) catch b = parseint(BigInt, b) end
end
tests = {[1, 34, 3, 98, 9, 76, 45, 4],
[54, 546, 548, 60], [1, 34, 3, 98, 9, 76, 45, 4, 54, 546, 548, 60]}
for t in tests
println("Maxconcating ", t) println(" ", maxconcat(t))
</lang>
- Output:
Maxconcating [1,34,3,98,9,76,45,4] 998764543431 Maxconcating [54,546,548,60] 6054854654 Maxconcating [1,34,3,98,9,76,45,4,54,546,548,60] 9987660548546544543431
Kotlin
<lang kotlin>import java.util.Comparator
val SORTER = Comparator<Int> { x, y ->
val xy = (x.toString() + y).toInt() val yx = (y.toString() + x).toInt() xy.compareTo(yx)
}
fun maxCat() {
fun findLargestSequence(array: IntArray): String { return array.sortedWith(SORTER).reversed().map { it.toString() }.joinToString("") } val source1 = intArrayOf(1, 34, 3, 98, 9, 76, 45, 4) println(findLargestSequence(source1))
val source2 = intArrayOf(54, 546, 548, 60); println(findLargestSequence(source2))
}</lang>
- Output:
998764543431 6054854654
Lua
<lang Lua>function icsort(numbers) table.sort(numbers,function(x,y) return (x..y) > (y..x) end) return numbers end
for _,numbers in pairs({{1, 34, 3, 98, 9, 76, 45, 4}, {54, 546, 548, 60}}) do print(('Numbers: {%s}\n Largest integer: %s'):format( table.concat(numbers,","),table.concat(icsort(numbers)) )) end</lang>
- Output:
Numbers: {1,34,3,98,9,76,45,4} Largest integer: 998764543431 Numbers: {54,546,548,60} Largest integer: 6054854654
Mathematica
<lang Mathematica>makeLargestInt[list_] := Module[{sortedlist},
sortedlist = Sort[list, Order[ToString[#1] <> ToString[#2], ToString[#2] <> ToString[#1]] < 0 &]; Map[ToString, sortedlist] // StringJoin // FromDigits ]
(* testing with two examples *) makeLargestInt[{1, 34, 3, 98, 9, 76, 45, 4}] makeLargestInt[{54, 546, 548, 60}]</lang>
- Output:
998764543431 6054854654
NetRexx
<lang NetRexx>/* NetRexx */ options replace format comments java crossref symbols nobinary
runSample(arg) return
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ method largestInt(il) public static
ri = wa = -- put the list into an indexed string wa[0] = il.words loop ww = 1 to wa[0] wa[ww] = il.word(ww) end ww
-- order the list loop wx = 1 to wa[0] - 1 loop wy = wx + 1 to wa[0] xx = wa[wx] yy = wa[wy] xy = xx || yy yx = yy || xx if xy < yx then do -- swap xx and yy wa[wx] = yy wa[wy] = xx end end wy end wx
-- rebuild list from indexed string loop ww = 1 to wa[0] ri = ri wa[ww] end ww return ri.space(0) -- concatenate the list elements into a single numeric
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ method runSample(arg) private static
ints = [ - '1 34 3 98 9 76 45 4', - '54 546 548 60' - ] loop il over ints say largestInt(il).right(20) ':' il.space(1, ',') end il return
</lang>
- Output:
998764543431 : 1,34,3,98,9,76,45,4 6054854654 : 54,546,548,60
Nim
<lang nim>import algorithm, sequtils, strutils, future
proc maxNum(x: seq[int]): string =
var c = x.mapIt(string, $it) c.sort((x, y) => cmp(y&x, x&y)) c.join()
echo maxNum(@[1, 34, 3, 98, 9, 76, 45, 4]) echo maxNum(@[54, 546, 548, 60])</lang>
- Output:
998764543431 6054854654
OCaml
<lang ocaml>let myCompare a b = compare (b ^ a) (a ^ b) let icsort nums = String.concat "" (List.sort myCompare (List.map string_of_int nums))</lang>
- testing
# icsort [1;34;3;98;9;76;45;4];; - : string = "998764543431" # icsort [54;546;548;60];; - : string = "6054854654"
Oforth
<lang Oforth>: largestInt map(#asString) sortWith(#[ 2dup + -rot swap + > ]) sum asInteger ;</lang>
- Output:
[ [1, 34, 3, 98, 9, 76, 45, 4], [54, 546, 548, 60] ] map(#largestInt) . [998764543431, 6054854654]
Pascal
tested with freepascal.Used a more extreme example 3.
algorithm 3
<lang pascal>const
base = 10; MaxDigitCnt = 11; source1 : array[0..7] of integer = (1, 34, 3, 98, 9, 76, 45, 4); source2 : array[0..3] of integer = (54,546,548,60); source3 : array[0..3] of integer = (60, 54,545454546,0);
type
tdata = record datOrg, datMod : LongWord; datStrOrg : string[MaxDigitCnt]; end; tArrData = array of tData;
procedure DigitCount(var n: tdata); begin
with n do //InttoStr is very fast str(datOrg,datStrOrg);
end;
procedure InsertData(var n: tdata;data:LongWord); begin
n.datOrg := data; DigitCount(n);
end;
function FindMaxLen(const ArrData:tArrData): LongWord; var
cnt : longInt; res,t : LongWord;
begin
res := 0;// 1 is minimum for cnt := High(ArrData) downto Low(ArrData) do begin t := length(ArrData[cnt].datStrOrg); IF res < t then res := t; end; FindMaxLen := res;
end;
procedure ExtendCount(var ArrData:tArrData;newLen: integer); var
cnt, i,k : integer;
begin
For cnt := High(ArrData) downto Low(ArrData) do with ArrData[cnt] do begin datMod := datOrg; i := newlen-length(datStrOrg); k := 1; while i > 0 do begin datMod := datMod *Base+Ord(datStrOrg[k])-Ord('0'); inc(k); IF k >length(datStrOrg) then k := 1; dec(i); end; end;
end;
procedure SortArrData(var ArrData:tArrData); var
i, j,idx : integer; tmpData : tData;
begin
For i := High(ArrData) downto Low(ArrData)+1 do begin idx := i; j := i-1; For j := j downto Low(ArrData) do IF ArrData[idx].datMod < ArrData[j].datMod then idx := j; IF idx <> i then begin tmpData := ArrData[idx]; ArrData[idx]:= ArrData[i]; ArrData[i] := tmpData; end; end;
end;
procedure ArrDataOutput(const ArrData:tArrData); var
i,l : integer; s : string;
begin { the easy way
For i := High(ArrData) downto Low(ArrData) do write(ArrData[i].datStrOrg); writeln; *} l := 0; For i := High(ArrData) downto Low(ArrData) do inc(l,length(ArrData[i].datStrOrg)); setlength(s,l); l:= 1; For i := High(ArrData) downto Low(ArrData) do with ArrData[i] do begin move(datStrOrg[1],s[l],length(datStrOrg)); inc(l,length(datStrOrg)); end; writeln(s);
end;
procedure HighestInt(var ArrData:tArrData); begin
ExtendCount(ArrData,FindMaxLen(ArrData)); SortArrData(ArrData); ArrDataOutput(ArrData);
end;
var
i : integer; tmpData : tArrData;
begin
// Source1 setlength(tmpData,length(source1)); For i := low(tmpData) to high(tmpData) do InsertData(tmpData[i],source1[i]); HighestInt(tmpData); // Source2 setlength(tmpData,length(source2)); For i := low(tmpData) to high(tmpData) do InsertData(tmpData[i],source2[i]); HighestInt(tmpData); // Source3 setlength(tmpData,length(source3)); For i := low(tmpData) to high(tmpData) do InsertData(tmpData[i],source3[i]); HighestInt(tmpData);
end.</lang>
- Output:
998764543431 6054854654 60545454546540
Inspired by Haskell
generate the repetition by dividing /(10^CountDigits-1) http://rosettacode.org/wiki/Largest_int_from_concatenated_ints#Compare_repeated_string_method
<lang pascal>const
base = 10; MaxDigitCnt = 11; source1 : array[0..7] of LongInt = (10 , 34, 3, 98, 9, 76, 45, 4); source2 : array[0..3] of LongInt = (54,546,548,60); source3 : array[0..3] of LongInt = (0,2121212122,21,60);
type
tdata = record datMod : double; datOrg : LongInt;
//InttoStr is very fast and the string is always needed
datStrOrg : string[MaxDigitCnt]; end; tArrData = array of tData;
procedure InsertData(var n: tdata;data:LongWord); begin
with n do begin datOrg := data; str(datOrg,datStrOrg); end;
end;
function FindMaxLen(const ArrData:tArrData): LongWord; var
cnt : longInt; res,t : LongWord;
begin
res := 0;// 1 is minimum for cnt := High(ArrData) downto Low(ArrData) do begin t := length(ArrData[cnt].datStrOrg); IF res < t then res := t; end; FindMaxLen := res;
end;
procedure ExtendData(var ArrData:tArrData;newLen: integer); var
cnt, i : integer;
begin
For cnt := High(ArrData) downto Low(ArrData) do with ArrData[cnt] do begin //generating 10^length(datStrOrg) datMod := 1; i := length(datStrOrg); // i always >= 1 repeat datMod := base*datMod; dec(i); until i <= 0;
// 1/(datMod-1.0) = 1/(9...9)
datMod := datOrg/(datMod-1.0)+datOrg; i := newlen-length(datStrOrg); For i := i downto 1 do datMod := datMod*Base; end;
end;
procedure SortArrData(var ArrData:tArrData); //selection sort var
i, j,idx : integer; tmpData : tData;
begin
For i := High(ArrData) downto Low(ArrData)+1 do begin idx := i; j := i-1; //select max For j := j downto Low(ArrData) do IF ArrData[idx].datMod < ArrData[j].datMod then idx := j; //finally swap IF idx <> i then begin tmpData := ArrData[idx]; ArrData[idx]:= ArrData[i]; ArrData[i] := tmpData; end; end;
end;
procedure ArrDataOutput(const ArrData:tArrData); var
i : integer;
begin { the easy way}
For i := High(ArrData) downto Low(ArrData) do write(ArrData[i].datStrOrg); writeln;
end;
procedure HighestInt(var ArrData:tArrData); begin
ExtendData(ArrData,FindMaxLen(ArrData)); SortArrData(ArrData); ArrDataOutput(ArrData);
end;
var
i : integer; tmpData : tArrData;
begin
// Source1 setlength(tmpData,length(source1)); For i := low(tmpData) to high(tmpData) do InsertData(tmpData[i],source1[i]); HighestInt(tmpData); // Source2 setlength(tmpData,length(source2)); For i := low(tmpData) to high(tmpData) do InsertData(tmpData[i],source2[i]); HighestInt(tmpData); // Source3 setlength(tmpData,length(source3)); For i := low(tmpData) to high(tmpData) do InsertData(tmpData[i],source3[i]); HighestInt(tmpData);
end.</lang>
- Output:
9987645434310 6054854654 602121212122210>
PARI/GP
Sorts then joins. Most of the noise comes from converting a vector of integers into a concatenated integer: eval(concat(apply(n->Str(n),v)))
. Note that the short form eval(concat(apply(Str,v)))
is not valid here because Str
is variadic.
<lang parigp>large(v)=eval(concat(apply(n->Str(n),vecsort(v,(x,y)->eval(Str(y,x,"-",x,y)))))); large([1, 34, 3, 98, 9, 76, 45, 4]) large([54, 546, 548, 60])</lang>
- Output:
%1 = 998764543431 %2 = 6054854654
Perl
<lang perl>sub maxnum {
join , sort { "$b$a" cmp "$a$b" } @_
}
print maxnum(1, 34, 3, 98, 9, 76, 45, 4), "\n"; print maxnum(54, 546, 548, 60), "\n";</lang>
- Output:
998764543431 6054854654
Perl 6
<lang perl6>sub maxnum(@x) {
[~] @x.sort: -> $a, $b { $b ~ $a leg $a ~ $b }
}
say maxnum <1 34 3 98 9 76 45 4>; say maxnum <54 546 548 60>;</lang>
- Output:
998764543431 6054854654
PHP
<lang php>function maxnum($nums) {
usort($nums, function ($x, $y) { return strcmp("$y$x", "$x$y"); }); return implode(, $nums);
}
echo maxnum(array(1, 34, 3, 98, 9, 76, 45, 4)), "\n"; echo maxnum(array(54, 546, 548, 60)), "\n";</lang>
- Output:
998764543431 6054854654
PicoLisp
Here are solutions for all three algorithms.
The third solution actually avoids padding the numbers, by converting them into circular lists and comparing these. As a drawback, however, this works only for unique lists (as the comparison of identical numbers would not terminate), so a better solution might involve additional checks. <lang PicoLisp>(load "@lib/simul.l") # For 'permute'</lang>
Algorithm 1
<lang PicoLisp>(for L '((1 34 3 98 9 76 45 4) (54 546 548 60))
(prinl (maxi format (permute L))) )</lang>
Algorithm 2
<lang PicoLisp>(for L '((1 34 3 98 9 76 45 4) (54 546 548 60))
(prinl (sort L '((A B) (> (format (pack A B)) (format (pack B A)) ) ) ) ) )</lang>
Algorithm 3
<lang PicoLisp>(for L '((1 34 3 98 9 76 45 4) (54 546 548 60))
(prinl (flip (by '((N) (apply circ (chop N))) sort L) ) ) )</lang>
- Output:
in all three cases
998764543431 6054854654
PL/I
<lang pli> /* Largest catenation of integers 16 October 2013 */ /* Sort using method 2, comparing pairs of adjacent integers. */
Largest: procedure options (main);
declare s(*) char (20) varying controlled, n fixed binary; get (n); allocate s(n); get list (s); s = trim(s); put skip edit (s) (a, x(1)); put skip list ('Largest integer=', Largest_integer());
largest_integer: procedure () returns (char(100) varying);
declare sorted bit (1); declare (true value ('1'b), false value ('0'b)) bit (1); declare i fixed binary; declare temp character(20) varying;
do until (sorted); sorted = true; do i = 1 to n-1; if char(s(i)) || char(s(i+1)) < char(s(i+1)) || char(s(i)) then do; temp = s(i); s(i) = s(i+1); s(i+1) = temp; sorted = false; end; end; end; return (string(s));
end largest_integer; end Largest; </lang>
54 546 548 60 Largest integer= 6054854654 1 34 3 98 9 76 45 4 Largest integer= 998764543431
Prolog
Works with SWI-Prolog 6.5.3.
All permutations method
<lang Prolog>largest_int_v1(In, Out) :- maplist(name, In, LC), aggregate(max(V), get_int(LC, V), Out).
get_int(LC, V) :-
permutation(LC, P),
append(P, LV),
name(V, LV).
</lang>
- Output:
?- largest_int_v1([1, 34, 3, 98, 9, 76, 45, 4], Out). Out = 998764543431. ?- largest_int_v1([54, 546, 548, 60], Out). Out = 6054854654.
Method 2
<lang Prolog>largest_int_v2(In, Out) :- maplist(name, In, LC), predsort(my_sort,LC, LCS), append(LCS, LC1), name(Out, LC1).
my_sort(R, L1, L2) :-
append(L1, L2, V1), name(I1, V1),
append(L2, L1, V2), name(I2, V2),
( I1 < I2, R = >; I1 = I2, R = '='; R = <).
% particular case 95 958 my_sort(>, [H1], [H1, H2 | _]) :- H1 > H2.
my_sort(<, [H1], [H1, H2 | _]) :- H1 < H2.
my_sort(R, [H1], [H1, H1 | T]) :- my_sort(R, [H1], [H1 | T]).
% particular case 958 95 my_sort(>, [H1, H2 | _], [H1]) :- H1 > H2.
my_sort(<, [H1, H2 | _], [H1]) :- H1 < H2.
my_sort(R, [H1, H1 | T], [H1]) :- my_sort(R, [H1 | T], [H1]) . </lang>
- Output:
?- largest_int_v2([1, 34, 3, 98, 9, 76, 45, 4], Out). Out = 998764543431 . ?- largest_int_v2([54, 546, 548, 60], Out). Out = 5486054654 .
Python
Python: Sort on comparison of concatenated ints method
This also shows one of the few times where cmp= is better than key= on sorted()
<lang python>try:
cmp # Python 2 OK or NameError in Python 3 def maxnum(x): return .join(sorted((str(n) for n in x), cmp=lambda x,y:cmp(y+x, x+y)))
except NameError:
# Python 3 from functools import cmp_to_key def cmp(x, y): return -1 if x<y else ( 0 if x==y else 1) def maxnum(x): return .join(sorted((str(n) for n in x), key=cmp_to_key(lambda x,y:cmp(y+x, x+y))))
for numbers in [(1, 34, 3, 98, 9, 76, 45, 4), (54, 546, 548, 60):
print('Numbers: %r\n Largest integer: %15s' % (numbers, maxnum(numbers)))</lang>
- Output:
Numbers: (1, 34, 3, 98, 9, 76, 45, 4) Largest integer: 998764543431 Numbers: (54, 546, 548, 60) Largest integer: 6054854654
Python: Compare repeated string method
<lang python>def maxnum(x):
maxlen = len(str(max(x))) return .join(sorted((str(v) for v in x), reverse=True, key=lambda i: i*(maxlen * 2 // len(i))))
for numbers in [(212, 21221), (1, 34, 3, 98, 9, 76, 45, 4), (54, 546, 548, 60)]:
print('Numbers: %r' % (numbers, maxnum(numbers)))</lang>
- Output:
Numbers: (212, 21221) Largest integer: 21221221 Numbers: (1, 34, 3, 98, 9, 76, 45, 4) Largest integer: 998764543431 Numbers: (54, 546, 548, 60) Largest integer: 6054854654
<lang python>from fractions import Fraction from math import log10
def maxnum(x):
return .join(str(n) for n in sorted(x, reverse=True, key=lambda i: Fraction(i, 10**(int(log10(i))+1)-1)))
for numbers in [(1, 34, 3, 98, 9, 76, 45, 4), (54, 546, 548, 60)]:
print('Numbers: %r\n Largest integer: %15s' % (numbers, maxnum(numbers)))</lang>
- Output as first Python example, above.
Python: Try all permutations method
<lang python>from itertools import permutations def maxnum(x):
return max(int(.join(n) for n in permutations(str(i) for i in x)))
for numbers in [(1, 34, 3, 98, 9, 76, 45, 4), (54, 546, 548, 60)]:
print('Numbers: %r\n Largest integer: %15s' % (numbers, maxnum(numbers)))</lang>
- Output as above.
Racket
<lang Racket>
- lang racket
(define (largest-int ns)
(string->number (apply ~a (sort ns (λ(x y) (string>? (~a x y) (~a y x)))))))
(map largest-int '((1 34 3 98 9 76 45 4) (54 546 548 60)))
- -> '(998764543431 6054854654)
</lang>
REXX
The algorithm used is based on exact comparisons (left to right) with right digit fill of the left digit.
This allows the integers to be of any size.
This REXX version works with any integer (negative, zero, positive), and does some basic error checking to
verify that the numbers are indeed integers (and it also normalizes the integers).
<lang rexx>/*REXX program constructs the largest integer from a list using concatenation. */
@.=.; @.1 = '{1, 34, 3, 98, 9, 76, 45, 4}' /*the 1st integer list to be used. */
@.2 = '{54, 546, 548, 60}' /* " 2nd " " " " " */ @.3 = '{ 4, 45, 54, 5}' /* " 3rd " " " " " */ /* [↓] process all the integer lists.*/ do j=1 while @.j\==.; $= /*keep truckin' until lists exhausted. */ z=space(translate(@.j, , '])},{([')) /*perform scrubbing on the integer list*/ _=length(space(z, 0)) + 2 /*determine the largest possibility. */ if _>digits() then numeric digits _ /*use enough decimal digits for maximum*/ /* [↓] examine each number in the list*/ do while z\==; index=1 /*keep examining the list until done.*/ big=isOK(word(z,1)) /*assume that first integer is biggest.*/ /* [↓] check the rest of the integers.*/ do k=2 to words(z); #=isOK(word(z,k)) /*obtain an an integer from the list. */ L=max(length(big), length(#)) /*get the maximum length of the integer*/ if left(#, L, left(#,1)) <<= left(big, L, left(big,1)) then iterate big=#; index=k /*we found a new biggie (and the index)*/ end /*k*/ /* [↑] find max concatenated integer. */
z=space(delword(z, index, 1)) /*delete this maximum integer from list*/ $=$ || big /*append " " " ───► $.*/ end /*while z*/ /* [↑] process all integers in a list.*/
say right($,digits()) ' max for: ' @.j /*show the maximum integer and the list*/ end /*j*/ /* [↑] process each list of integers. */
exit /*stick a fork in it, we're all done. */ /*──────────────────────────────────────────────────────────────────────────────────────*/ isOK: parse arg ?; if datatype(?,'W') then return abs(?)/1 /*normalize the integer.*/
say; say '***error*** number ' ? "isn't an integer."; say; exit 13</lang>
output using the default internal integer lists:
998764543431 max for: {1, 34, 3, 98, 9, 76, 45, 4} 6054854654 max for: {54, 546, 548, 60} 554454 max for: { 4, 45, 54, 5}
Ring
<lang ring> nums=[1,34,3,98,9,76,45,4] see largestInt(8) + nl nums=[54,546,548,60] see largestInt(4) + nl
func largestInt len l = "" sorted = false while not sorted
sorted=true for i=1 to len-1 a=string(nums[i]) b=string(nums[i+1]) if a+b<b+a temp = nums[i] nums[i] = nums[i+1] nums[i+1] = temp sorted=false ok next
end for i=1 to len
l+=string(nums[i])
next return l </lang> Output:
998764543431 6054854654
Ruby
Sort on comparison of concatenated ints method
<lang Ruby>def icsort nums
nums.sort { |x, y| "#{y}#{x}" <=> "#{x}#{y}" }
end
[[54, 546, 548, 60], [1, 34, 3, 98, 9, 76, 45, 4]].each do |c|
p c # prints nicer in Ruby 1.8 puts icsort(c).join
end</lang>
- Output:
[54, 546, 548, 60] 6054854654 [1, 34, 3, 98, 9, 76, 45, 4] 998764543431
Compare repeated string method
<lang ruby>def icsort nums
maxlen = nums.max.to_s.length nums.map{ |x| x.to_s }.sort_by { |x| x * (maxlen * 2 / x.length) }.reverse
end
[[54, 546, 548, 60], [1, 34, 3, 98, 9, 76, 45, 4]].each do |c|
p c # prints nicer in Ruby 1.8 puts icsort(c).join
end</lang>
- Output as above.
<lang ruby>require 'rational' #Only needed in Ruby < 1.9
def icsort nums
nums.sort_by { |i| Rational(i, 10**(Math.log10(i).to_i+1)-1) }.reverse
end
[[54, 546, 548, 60], [1, 34, 3, 98, 9, 76, 45, 4]].each do |c|
p c # prints nicer in Ruby 1.8 puts icsort(c).join
end</lang>
- Output as above.
Run BASIC
<lang runbasic>a1$ = "1, 34, 3, 98, 9, 76, 45, 4" a2$ = "54,546,548,60"
print "Max Num ";a1$;" = ";maxNum$(a1$) print "Max Num ";a2$;" = ";maxNum$(a2$)
function maxNum$(a1$) while word$(a1$,i+1,",") <> ""
i = i + 1 a$(i) = trim$(word$(a1$,i,","))
wend
s = 1 while s = 1
s = 0 for j = 1 to i -1 if a$(j)+a$(j+1) < a$(j+1)+a$(j) then h$ = a$(j) a$(j) = a$(j+1) a$(j+1) = h$ s = 1 end if next j
wend
for j = 1 to i
maxNum$ = maxNum$ ; a$(j)
next j end function</lang>
- Output:
Max Num 1, 34, 3, 98, 9, 76, 45, 4 = 998764543431 Max Num 54,546,548,60 = 6054854654
Scala
<lang Scala>object LIFCI extends App {
def lifci(list: List[Long]) = list.permutations.map(_.mkString).max
println(lifci(List(1, 34, 3, 98, 9, 76, 45, 4))) println(lifci(List(54, 546, 548, 60)))
}</lang>
- Output:
998764543431 6054854654
Scheme
<lang Scheme>(define (cat . nums) (apply string-append (map number->string nums)))
(define (my-compare a b) (string>? (cat a b) (cat b a)))
(map (lambda (xs) (string->number (apply cat (sort xs my-compare))))
'((1 34 3 98 9 76 45 4) (54 546 548 60)))</lang>
- Output:
(998764543431 6054854654)
Sidef
<lang ruby>func maxnum(nums) {
nums.sort {|x,y| "#{y}#{x}" <=> "#{x}#{y}" };
}
[[54, 546, 548, 60], [1, 34, 3, 98, 9, 76, 45, 4]].each { |c|
say maxnum(c).join.to_num;
}</lang>
- Output:
6054854654 998764543431
Tcl
<lang tcl>proc intcatsort {nums} {
lsort -command {apply {{x y} {expr {"$y$x" - "$x$y"}}}} $nums
}</lang> Demonstrating: <lang tcl>foreach collection {
{1 34 3 98 9 76 45 4} {54 546 548 60}
} {
set sorted [intcatsort $collection] puts "\[$collection\] => \[$sorted\] (concatenated: [join $sorted ""])"
}</lang>
- Output:
[1 34 3 98 9 76 45 4] => [9 98 76 45 4 34 3 1] (concatenated: 998764543431) [54 546 548 60] => [60 548 546 54] (concatenated: 6054854654)
VBScript
<lang vb> Function largestint(list) nums = Split(list,",") Do Until IsSorted = True IsSorted = True For i = 0 To UBound(nums) If i <> UBound(nums) Then a = nums(i) b = nums(i+1) If CLng(a&b) < CLng(b&a) Then tmpnum = nums(i) nums(i) = nums(i+1) nums(i+1) = tmpnum IsSorted = False End If End If Next Loop For j = 0 To UBound(nums) largestint = largestint & nums(j) Next End Function
WScript.StdOut.Write largestint(WScript.Arguments(0)) WScript.StdOut.WriteLine </lang>
- Output:
F:\>cscript /nologo largestint.vbs 1,34,3,98,9,76,45,4 998764543431 F:\>cscript /nologo largestint.vbs 54,546,548,60 6054854654
Vim Script
This solution is intended to be run as an Ex command within a buffer containing the integers to be processed, one per line. <lang Vim>%s/\(.\+\)/\1\1/ | sort! | %s/\(.\+\)\1\n/\1/</lang>
- Demonstration
<lang Bash>$ paste -s nums 1 34 3 98 9 76 45 4 $ vim -S icsort.vim nums 998764543431</lang>
zkl
<lang zkl>fcn bigCI(ns){
ns.apply("toString").sort(fcn(a,b){ (a+b)>(b+a) }).concat();
}</lang> <lang zkl>bigCI(T(1, 34, 3, 98, 9, 76, 45, 4)).println(); bigCI(T(54, 546, 548, 60)).println();</lang>
- Output:
998764543431 6054854654
- Programming Tasks
- Solutions by Programming Task
- Ada
- AutoHotkey
- AWK
- BBC BASIC
- Bracmat
- C
- C++
- C sharp
- Ceylon
- Clojure
- Common Lisp
- D
- Elixir
- Erlang
- Fortran
- Go
- Groovy
- Haskell
- Icon
- Unicon
- J
- Java
- JavaScript
- Jq
- Julia
- Kotlin
- Lua
- Mathematica
- NetRexx
- Nim
- OCaml
- Oforth
- Pascal
- PARI/GP
- Perl
- Perl 6
- PHP
- PicoLisp
- PL/I
- Prolog
- Python
- Racket
- REXX
- Ring
- Ruby
- Run BASIC
- Scala
- Scheme
- Sidef
- Tcl
- VBScript
- Vim Script
- Zkl