前書き

定義上、一意の識別子は一意である必要があります。同じ識別子が複数あると、予期しないデータが取得されます。ただし、複数のソースから同時に到着するデータでは、一意性を確保するのは困難です。識別子のリストを一意化する関数を作成します。

これは恐らく私が書いた最悪のパズルの毛羽ですが、あなたはそのアイデアを得ます。

必要条件

0個以上の正の整数のリストが与えられた場合、最初から最後まで次の規則を各数値に適用します。

• 番号がその種の最初のものである場合、それを保持します。
• 以前に番号に遭遇した場合は、入力リスト全体または既存の出力のどこにも見られない正の最小整数に置き換えます。

ソリューションの場合：

• ソリューションは、プログラムまたは関数である可能性があります。
• 入力は、引数として渡される文字列、配列、またはキーボード入力です。
• 出力は、文字列、配列、または画面に出力される場合があります。
• 出力リストのすべての番号は異なります。

仮定

• 入力リストはきれいです。正の整数のみが含まれます。
• 正の整数の範囲は1〜2 ³¹ -1です。
• プログラムの変数に使用できるメモリは256 MB未満です。（基本的に、2,147,483,648要素の配列は許可されていません。）

テストケース

Input:  empty
Output: empty

Input:  5
Output: 5

Input:  1, 4, 2, 5, 3, 6
Output: 1, 4, 2, 5, 3, 6

Input:  3, 3, 3, 3, 3, 3
Output: 3, 1, 2, 4, 5, 6

Input:  6, 6, 4, 4, 2, 2
Output: 6, 1, 4, 3, 2, 5

Input:  2147483647, 2, 2147483647, 2
Output: 2147483647, 2, 1, 3

得点

シンプルなコードゴルフ。来週のこの時点までの最低バイト数が勝ちです。

Brachylog, 8 bytes

{|∧ℕ₁}ᵐ≠

Try it online!

Explanation

{    }ᵐ     Map on the Input:
              Input = Output…
 |            …or…
  ∧ℕ₁         …Output is in [1,+inf)
       ≠    All elements of the output must be different
            (Implicit labeling)

Java 8, 158 144 bytes

a->{int m=0;String r="",c=",",b=r;for(int x:a)b+=x+c;for(int x:a)if(r.contains(x+c)){for(;(r+b).contains(++m+c););r+=m+c;}else r+=x+c;return r;}

• .contains(m+c);m++) to .contains(++m+c);) to save 1 byte, and simultaneously converted to Java 8 to save 13 more bytes.

Explanations:

Try it here.

a->{                      // Method with integer-array parameter and String return-type
  int m=0;                //  Lowest integer
  String r="",            //  Result-String
         c=",",           //  Comma delimiter for result String
         b=r;for(int x:a)b+=x+c;
                          //  Input array as String
  for(int x:a)            //  Loop (2) over the integers in the array
    if(r.contains(x+c)){  //   If the result already contains this integer
      for(;(r+b).contains(++m+c););
                          //    Inner (3) as long as either the result-String or array-String contains the lowest integer
                          //     and raise the lowest integer before every iteration by 1
      r+=m+c;             //    Append the result with this lowest not-present integer
    }else                 //   Else:
      r+=x+c;             //    Append the result-String with the current integer
                          //  End of loop (2) (implicit / single-line body)
  return r;               //  Return the result-String
}                         // End of method

JavaScript (ES6), 49 bytes

a=>a.map(g=(e,i)=>a.indexOf(e)-i?g(++n,-1):e,n=0)

Ruby, 63 bytes

->a{r=*0..a.size;r.map{|i|[a[i]]-a[0,i]==[]?a[i]=(r-a)[1]:0};a}

Try it online!

Explanation

->a{                                    # Anonymous function with one argument
    r=*0..a.size;                       # Numbers from 0 to array size
    r.map{|i|                           # For all numbers in range:
        [a[i]]                          #  Get array with just a[i]
              -a[0,i]                   #  Remove elements from array that are
                                        #    also in a[0..i-1]
                    ==[]?               #  Check if result is an empty array
                        a[i]=           #  If true, set a[i] to:
                             (r-a)      #   Remove elements from number range
                                        #     that are also in input array
                                  [1]   #   Get second element (first non-zero)
                        :0};            #  If false, no-op
                            a}          # Return modified array

05AB1E, 17 16 18 bytes

vy¯yåi¹gL¯K¹K¬}ˆ}¯

Try it online!

Explanation

v                    # for each y in input
 y                   # push y
  ¯yåi               # if y exist in global list
      ¹gL            # push [1 ... len(input)]
         ¯K          # remove any number that is already in global list
           ¹K        # remove any number that is in the input
             ¬       # get the first (smallest)
              }      # end if
               ˆ     # add to global list
                }¯   # end loop, push and output global list

PHP, 121 Bytes

<?$n=array_diff(range(0,count($g=$_GET)),$g);sort($n);$r=[];foreach($g as$v)$r[]=in_array($v,$r)?$n[++$i]:$v;print_r($r);

Online Version

Expanded

$n=array_diff(range(0,count($g=$_GET)),$g); # create array of ascending values which are not in input array plus zero
sort($n); # minimize keys
$r=[];  # empty result array
foreach($g as$v) # loop input array
  $r[]=in_array($v,$r)?$n[++$i]:$v; # if value is not in result array add value else take new unique value skip zero through ++$i
print_r($r); # output result array

Python 2, 77 79 bytes

a=input();u=[];j=1
for x in a:
 u+=[[x,j][x in u]]
 while j in u+a:j+=1
print u

Takes keyboard input, like [3, 3, 3, 3, 3, 3].

Just keep track of the smallest positive integer j not used so far. For each element x of the input, output x if x hasn't already been used, otherwise output j. Finally, update j each time you output something.

EDITED: to fix a mistake handling input of [6, 6, 4, 4, 2, 2]. Thanks to @Rod for pointing out the mistake as well as a fix. The mistake was that, in the event of a duplicate entry, it would output the smallest number unused to that point in the list, even if that output appeared later in the input. (This was wrong, as clarified in the post and the comments, but I still messed it up somehow.) Anyway, the fix was to simply add the input list, a, to the set of values that could not be output in that case.

Haskell, 79 76 bytes

EDIT:

• -3 bytes: @nimi saw that head could be replaced by a pattern match.

([]#) is an anonymous function taking and returning a list. Use like ([]#)[2147483647, 2, 2147483647, 2].

(?)=notElem
s#(x:y)|z:_<-[x|x?s]++filter(?(s++y))[1..]=z:(z:s)#y
_#n=n
([]#)

Try it online!

How it works

• ? is an abbreviated operator for checking if an element is absent from a list.
• s#l handles the list of integers l, given a list s of integers already used.
  • x is the next integer to look at, y the remaining ones.
  • z is the integer chosen for the next spot. It's x if x is not an element of s, and the first positive integer neither in s nor in y otherwise.
  • (z:s)#y then recurses with z added to the used integer list.
  • n is an empty list, since nonempty lists have been handled in the previous line.
• The main function ([]#) takes a list, and calls # with it as second argument, and an empty list for the first argument.

APL (Dyalog 16.0), 34 bytes

(s↑v~⍨⍳⌈/v+s←+/b←(⍳≢v)≠⍳⍨v)@{b}v←⎕

Pyth, 21 20 bytes

.b?}NPY@-UhlQQ=hZNQ._
m?}edPd@-SlQQ~hZed._

Test suite

I'll add an explanation when I have the time.

C#, 135 bytes

Golfed

(int[] i)=>{for(int x=1,v,m=0,l=i.Length,y;x<l;x++){v=i[x];for(y=0;y<l&&v==i[x]?y<x:y<l;y++)if(i[y]==v){v=++m;y=-1;}i[x]=v;}return i;};

Ungolfed

( int[] i ) => {
   for( int x = 1, v, m = 0, l = i.Length, y; x < l; x++ ) {
      v = i[ x ];

      for( y = 0; y < l && v == i[ x ] ? y < x : y < l ; y++ )
         if( i[ y ] == v ) {
            v = ++m;
            y = -1;
         }

      i[ x ] = v;
   }

   return i;
};

Ungolfed readable

// Takes an array of Int32 objects ( 32-byte signed integers )
( int[] i ) => {

   // Cycles through each element on the array
   //   x: Scan position, starts at the 2nd element
   //   v: Value being processed
   //   m: The next minimum value to replace
   //   l: Size of the array, to save some byte count
   for( int x = 1, v, m = 0, l = i.Length, y; x < l; x++ ) {

      // Hold the value
      v = i[ x ];

      // Re-scan the array for a duplicate value up the the current position ( 'x' ) IF
      //   ... the currently hold value hasn't been modified
      //   ... otherwise, re-scans the entire array to find a suitable value to replace
      for( y = 0; y < l && v == i[ x ] ? y < x : y < l ; y++ )

         // Check if 'v' shares the same value with other element
         if( i[ y ] == v ) {

            // Set 'v' to the minimum value possible
            v = ++m;

            // Reset the scan position to validate the new value
            y = -1;
         }

      // Set the 'v' to the array
      i[ x ] = v;
   }

   // Return the array
   return i;
};

Full code

using System;
using System.Collections.Generic;

namespace Namespace {
   class Program {
      static void Main( String[] args ) {
         Func<Int32[], Int32[]> f = ( int[] i ) => {
            for( int x = 1, v, m = 0, l = i.Length, y; x < l; x++ ) {
               v = i[ x ];

               for( y = 0; y < l && v == i[ x ] ? y < x : y < l ; y++ )
                  if( i[ y ] == v ) {
                     v = ++m;
                     y = -1;
                  }

               i[ x ] = v;
            }

            return i;
         };

         List<Int32[]>
            testCases = new List<Int32[]>() {
               new Int32[] { },
               new Int32[] { 5 },
               new Int32[] { 1, 4, 2, 5, 3, 6 },
               new Int32[] { 3, 3, 3, 3, 3, 3 },
               new Int32[] { 6, 6, 4, 4, 2, 2 },
               new Int32[] { 2147483647, 2, 2147483647, 2 },
            };

         foreach( Int32[] testCase in testCases ) {
            Console.WriteLine( $" Input: {String.Join( ",", testCase )}\nOutput: {string.Join( ",", f( testCase ) )}\n" );
         }

         Console.ReadLine();
      }
   }
}

Releases

• v1.0 - 135 bytes - Initial solution.

Notes

• None

Python 2, 101 bytes

x=input()
for i,n in enumerate(x):x[i]=[n,list(set(range(1,len(x)+9))-set(x))[0]][n in x[:i]]
print x

Try it online! or Try all test cases

R, 39 46 bytes

Creates a vector from the input, then replaces the duplicated values with a range from 1 to a million that has the input values removed. Returns a numeric vector. No input will return the empty vector numeric(0).

i[duplicated(i)]=(1:1e6)[-(i=scan())];i

Try it online!

This will throw a warning about the length of the replacement vector

                           i=scan()     # set i as input
                 (1:1e6)                # 1 to a million (could go higher)
                 (1:1e6)[-(i=scan())]   # without input values
  duplicated(i)                         # duplicate values in i
i[duplicated(i)]=(1:1e6)[-(i=scan())]   # set duplicate i values to reduced range vector
                                     ;i # return result

C, 169 bytes 133 bytes

input = array a, output = modified array a

i=1,j,k,l;void f(int*a,int n){for(;i<n;i++)for(j=i-1;j>=0;j--)if(a[i]==a[j]){l=1;for(k=0;k<n;)if(l==a[k])k=l++?0:0;else k++;a[i]=l;}}

formatted

int i, j, k, l;
void f(int* a, int n)
{
    for (i = 1; i<n; i++)
        for (j = i - 1; j >= 0; j--)
            if (a[i] == a[j])
            {
                l = 1;
                for (k = 0; k<n;)
                    if (l == a[k])
                        k = l++ ? 0 : 0;
                    else
                        k++;
                a[i] = l;
            }
}

Too many bytes wasted for these loop. Does anybody think of shorten the code by inventing a new algorithm (which use less loop)? I was thinking but havent found one.

C# 7, 116 bytes

int[]f(int[]c){int j=0;int h()=>c.Contains(++j)?h():j;return c.Select((e,i)=>Array.IndexOf(c,e)<i?h():e).ToArray();}

Indented

int[] f(int[] c)
{
    int j = 0;
    int h() => c.Contains(++j) ? h() : j;
    return c
        .Select((e, i) => Array.IndexOf(c, e) < i ? h() : e)
        .ToArray();
}

Explained

• first occurrence of a number is always left as-is
• consecutive occurrences of a number are drawn from [1, 2, 3, ...], skipping values present in the input.

Online Version

Clojure, 72 bytes

#(reduce(fn[r i](conj r(if((set r)i)(nth(remove(set r)(range))1)i)))[]%)

A basic reduction. If i is contained in the output list so far, we'll take the 2nd element (1 when 0-indexed) from the infinite list of integers (range) from which we have removed those numbers that have already been used. Range starts from zero so we cannot take the first element but the second.

R, 74 bytes

reads the list from stdin; returns NULL for an empty input.

o=c();for(i in n<-scan())o=c(o,`if`(i%in%o,setdiff(1:length(n),o)[1],i));o

explanation:

o=c()                                #initialize empty list of outputs
for(i in n<-scan())                  # loop through the list after reading it from stdin
    o=c(o,                           # set the output to be the concatenation of o and
      `if`(i%in%o,                   # if we've seen the element before
           setdiff(1:length(n),o)[1] # the first element not in 1,2,...
           ,i))                      # otherwise the element
o                                    # print the output

1:length(n) may be used since we are guaranteed to never need a replacement from outside that range.

Try it online!

Axiom, 169 bytes

f a==(r:List INT:=[];for i in 1..#a repeat(~member?(a.i,r)=>(r:=concat(r,a.i));for j in 1..repeat if~member?(j,r)and(~member?(j,a)or j=a.i)then(r:=concat(r,j);break));r)

ungolf and result

ff(a)==
  r:List INT:=[]
  for i in 1..#a repeat
      ~member?(a.i,r)=>(r:=concat(r,a.i))
      for j in 1.. repeat
            if~member?(j,r)and(~member?(j,a) or j=a.i)then(r:=concat(r,j);break)
  r

(3) -> f([])
   (3)  []
                                                       Type: List Integer
(4) -> f([5])
   (4)  [5]
                                                       Type: List Integer
(5) -> f([1,4,2,5,3,6])
   (5)  [1,4,2,5,3,6]
                                                       Type: List Integer
(6) -> f([3,3,3,3,3,3])
   (6)  [3,1,2,4,5,6]
                                                       Type: List Integer
(7) -> f([6, 6, 4, 4, 2, 2])
   (7)  [6,1,4,3,2,5]
                                                       Type: List Integer
(8) -> f([2147483647, 2, 2147483647, 2])
   (8)  [2147483647,2,1,3]
                                                       Type: List Integer