# Randomize the scalars of an array

You must fill an array with every number from 0-n inclusive. No numbers should repeat. However they must be in a random order.

# Rules

All standard rules and standard loopholes are banned

The array must be generated pseudo-randomly. Every possible permutation should have a equal probability.

# Input

n in any way allowed in the I/O post on meta.

# Output

The array of numbers scrambled from 0-n inclusive.

• the output can be separated by newlines? – DrnglVrgs May 30 '17 at 20:48
• @Riley opps that was meant to be gone sorry. – Christopher May 31 '17 at 0:48
• @DrnglVrgs yes it can – Christopher May 31 '17 at 0:48
• By "numbers" I assume you mean "integers"? – Zacharý Aug 10 '17 at 1:12
• @KevinCruijssen IMO lists = array but with searching support. So sure use a list – Christopher Aug 11 '17 at 14:59

# Perl 6, 14 bytes

{pick *,0..$_}  Try it ## Expanded: { # bare block lambda with implicit parameter ｢$_｣

pick      # choose randomly without repeats
*,      # Whatever (all)
0 .. $_ # Range from 0, to the input (inclusive) }  # 05AB1E, 3 bytes Ý.r  Try it online! Ý # Make a list from 0 to input .r # Shuffle it randomly  # Pyth, 3 bytes .Sh  Demonstration .S is shuffle. It implicitly casts an input integer n to the range [0, 1, ..., n-1]. h is +1, and the input is taken implicitly. # R, 16 bytes sample(0:scan())  reads from stdin. sample randomly samples from the input vector, returning a (pseudo)random sequence. Try it online! # Jelly, 3 bytes 0rẊ  Try it online! Explanaion: 0rẊ 0r Inclusive range 0 to input. Ẋ Shuffle. Implicit print.  ## Alternate solution, 3 bytes ‘ḶẊ  Explanation: ‘ḶẊ ‘ Input +1 Ḷ Range 0 to argument. Ẋ Shuffle.  Try it online! # Python 2, 51 bytes lambda n:sample(range(n+1),n+1) from random import*  Try it online! There is random.shuffle() but it modifies the argument in place instead of returning it... • You can use random.shuffle – caird coinheringaahing Apr 9 '18 at 17:01 • @cairdcoinheringaahing Yeah, but that wouldn't work. For example, lambda n:shuffle(range(n+1)) wouldn't write the output anywhere. – totallyhuman Apr 9 '18 at 19:18 # PHP, 42 Bytes $r=range(0,$argn);shuffle($r);print_r($r);  Try it online! # Bash, 18 11 bytes shuf -i0-$1


Try it online!

• If newlines are allowed we can forget the echo part – DrnglVrgs May 30 '17 at 20:53

# Mathematica, 24 bytes

RandomSample@Range[0,#]&


# MATL, 4 bytes

QZ@q


Try it online!

### Explanation

Q     % Implicitly input n. Add 1
Z@    % Random permutation of [1 2 ... n+1]
q     % Subtract 1, element-wise. Implicitly display


# Brachylog, 2 bytes

⟦ṣ


Try it online!

### Explanation

⟦     Range from 0 to Input
ṣ    Shuffle


# Japt, 4 bytes

ò öx


Try it online

    :Implicit input of integer U
ò   :Generate array of 0 to U.
öx  :Generate random permutation of array.
:Implicit output of result.

• Gosh darn it, I thought öx would be enough until I noticed the "inclusive" part. (You could replace the x with almost anything else, btw) – ETHproductions May 30 '17 at 20:38
• @ETHproductions, that was my first thought too. – Shaggy May 30 '17 at 20:55

# C#, 76 bytes

using System.Linq;i=>new int[i+1].Select(x=>i--).OrderBy(x=>Guid.NewGuid());


This returns an IOrderedEnumerable, I hope that's okay, or else I need a few more bytes for a .ToArray()

# CJam, 7 6 bytes

1 byte removed thanks to Erik the Outgolfer.

{),mr}


This is an anonymous block (function) that takes an integer from the stack and replaces it with the result. Try it online!

### Explanation

{     e# Begin block
)     e# Increment: n+1
,     e# Range: [0 1 ... n]
mr    e# Shuffle
}     e# End block

• Isn't {),mr} 1 byte shorter? – Erik the Outgolfer May 31 '17 at 18:43
• @EriktheOutgolfer Indeed! Thanks – Luis Mendo May 31 '17 at 21:19

# Java 8, 114111 97 bytes

import java.util.*;n->{List l=new Stack();for(;n>=0;l.add(n--));Collections.shuffle(l);return l;}


-3 bytes and bug-fixed thanks to @OlivierGrégoire.
-4 bytes thanks to @Jakob.
-10 bytes by removing .toArray().

Explanation:

Try it here.

import java.util.*;        // Required import for List, Stack and Collections
n->{                       // Method with integer parameter and Object-array return-type
List l=new Stack();      //  Initialize a List
for(;n>=0;l.add(n--));   //  Loop to fill the list with 0 through n
Collections.shuffle(l);  //  Randomly shuffle the List
return l;                //  Convert the List to an Object-array and return it
}                          // End of method

• Bug: doesn't include n. Fix and golf: for(n++;--n>=0;l.add(n));. Also, I say you don't need to return an array. Array and list are the same in most language, so just return the list. – Olivier Grégoire Aug 10 '17 at 13:23
• @OlivierGrégoire Woops.. That's what you get for not checking properly and just posting.. Thanks for the bug-fix (and 4 bytes saved in the process). – Kevin Cruijssen Aug 10 '17 at 13:27
• Well, three actually, because I edited again, having myself introduced another bug: > should be >=. – Olivier Grégoire Aug 10 '17 at 13:28
• -4 bytes: use a Stack instead of a Vector and change your loop to for(;n>=0;l.add(n--));. And returning a java.util.List is definitely fine. – Jakob Aug 10 '17 at 14:10

# Ohm, 2 bytes

#╟


Try it online!

# Pyth, 4 Bytes

.S}0


Try it here!

• You can golf to 3 bytes. .S with an integer argument is the same as .SU, and [0..n] can be coded as Uh, so you can use .SUh, which then becomes .Sh. – Erik the Outgolfer May 31 '17 at 18:46
• @EriktheOutgolfer thanks for the hint, but as someone has alread posted the solution you propose I will leave this as this. – KarlKastor May 31 '17 at 21:32
• Well, it's borderline whether that should've been a separate answer or not, but I believe it counts as a dupe, so even it being allowed, I'd consider it just builtin substitution, so nah, I didn't want to post separate, but isaacg did. – Erik the Outgolfer Jun 1 '17 at 9:30

# C, 75 bytes

a[99],z,y;f(n){if(n){a[n]=--n;f(n);z=a[n];a[n]=a[y=rand()%(n+1)];a[y]=z;}}


Recursive function that initializes from the array's end on the way in, and swaps with a random element before it on the way out.

• What if n > 98? – LegionMammal978 May 31 '17 at 15:11
• It would fail, of course, but input range wasn't specified in the problem. Please don't make me malloc :) – Computronium May 31 '17 at 15:40
• change a into a para to fit the rule more? – l4m2 Apr 9 '18 at 16:00
• 67 bytes – ceilingcat Apr 1 '19 at 3:57

# Charcoal, 33 bytes

Ａ…·⁰ＮβＦβ«ＡβδＡ‽δθＰＩθ↓Ａ⟦⟧βＦδ¿⁻θκ⊞βκ


Try it online! Link is to verbose version of code.

Apparently it takes 17 bytes to remove an element from a list in Charcoal.

Edit: These days it only takes three bytes, assuming you want to remove all occurrences of the item from the list. This plus other Charcoal changes cut the answer down to 21 bytes: Try it online!

• Yikes that is a lot – Christopher Jun 1 '17 at 1:16

# APL (Dyalog), 5 bytes

?⍨1+⊢


Try it online!

Assumes ⎕IO←0, which is default on many machines.

### Explanation

⊢ the right argument

1+ add 1 to it

?⍨ generate numbers 0 .. 1+⊢-1 and randomly deal them in an array so that no two numbers repeat

# q/kdb+, 11 bytes

Solution:

{(0-x)?1+x}


Example:

q){(0-x)?1+x}10
5 9 7 1 2 4 8 0 3 10
q){(0-x)?1+x}10
6 10 2 8 4 5 9 0 7 3
q){(0-x)?1+x}10
9 6 4 1 10 8 2 7 0 5


Explanation:

Use the ? operator with a negative input to give the full list of 0->n without duplicates:

{(0-x)?1+x} / solution
{         } / lambda expression
x  / implicit input
?     / rand
(0-x)      / negate x, 'dont put item back in the bag'


# TI-83 BASIC, 5 bytes (boring)

randIntNoRep(0,Ans


Yep, a builtin. randIntNoRep( is a two-byte token, and Ans is one byte.

### More fun, 34 bytes:

Ans→N
seq(X,X,0,N→L₁
rand(N+1→L₂
SortA(L₂,L₁
L₁


Straight from tibasicdev. Probably golfable, but I haven't found anything yet.

What this does: Sorts a random array, moving elements of the second arg (L₁ here) in the same way as their corresponding elements.

# JavaScript (ES6), 51 bytes

n=>[...Array(n+1).keys()].sort(_=>.5-Math.random())

• I don't think this is uniform; I've tried f(5) 10 times and 5 has been one of the last two items every time. – ETHproductions May 30 '17 at 21:47
• Just ran it again a couple of times myself and got 1,5,4,0,2,3 & 1,0,2,5,3,4. EDIT: And a few more prnt.sc/fe0goe – Shaggy May 30 '17 at 22:27
• Just ran a quick test which runs f(5) 1e5 times and finds the average position of each number in the results. The resulting array was [ 1.42791, 1.43701, 2.00557, 2.6979, 3.3993, 4.03231 ], so I don't think it's uniform. (code) – ETHproductions May 30 '17 at 22:35
• I think I have a 93 byte solution that could work. n=>(a=[...Array(n).keys(),n++]).reduce((a,v,i)=>([a[i],a[j]]=[a[j=n*Math.random()|0],v],a),a)? – kamoroso94 May 31 '17 at 7:39
• Sorting on the result of random() isn't uniform. See (for example) en.wikipedia.org/wiki/BrowserChoice.eu#Criticism – Neil May 31 '17 at 9:07

# Aceto, 1514 16 bytes

@lXp
Y!n
zi&
0r


Push zero on the stack, read an integer, construct a range and shuffle it:

Y
zi
0r


Set a catch mark, test length for 0, and (in that case) exit:

@lX
!


Else print the value, a newline, and jump back to the length test:

   p
n
&


(I had to change the code because I realized I misread the question and had constructed a range from 1-n, not 0-n.)

# Go, 92 bytes

Mostly losing to the need to seed the PRNG.

import(."fmt";."math/rand";."time")
func f(n int){Seed(Now().UnixNano());Println(Perm(n+1))}


Try it online!

# Ruby, 20 bytes

->n{[*0..n].shuffle}

# 8th, 4236 34 bytes

Code

>r [] ' a:push 0 r> loop a:shuffle


SED (Stack Effect Diagram) is n -- a

Usage and example

ok> 5 >r [] ' a:push 0 r> loop a:shuffle .
[2,5,0,3,1,4]


# Javascript (ES6), 68 bytes

n=>[...Array(n+1)].map((n,i)=>[Math.random(),i]).sort().map(n=>n[1])


Creates an array of form

[[Math.random(), 0],
[Math.random(), 1],
[Math.random(), 2],...]


Then sorts it and returns the last elements in the new order

# J, 11 Bytes

(?@!A.i.)>:


### Explanation:

         >:   | Increment
(?@!A.i.)     | Fork, (f g h) n is evaluated as (f n) g (h n)
i.      | Integers in range [0,n) inclusive
?@!          | Random integer in the range [0, n!)
A.        | Permute right argument according to left


Examples:

    0 A. i.>:5
0 1 2 3 4 5
1 A. i.>:5
0 1 2 3 5 4
(?@!A.i.)>: 5
2 3 5 1 0 4
(?@!A.i.)>: 5
0 3 5 1 2 4


proc R n {lsort -c {try expr\ rand()>.5 on 9} [if [incr n -1]+2 {concat [R $n] [incr n]}]}  Try it online! # Tcl, 96 bytes proc R n {proc f a\ b {expr rand()>.5} set i -1 while \$i<$n {lappend L [incr i]} lsort -c f$L}


Try it online!

• Try to outgolf got the same byte count: tio.run/… – sergiol Mar 20 '18 at 2:01