It's become somewhat of a tradition in PPCG that some users temporarily change their names by an anagram (a new name formed by reordering the letters of the old).

Sometimes it gets difficult to find out who is who. I could use a program or function to tell if two phrases are anagrams of each other.

The challenge

The program or function should take two strings and produce a truthy result if they are anagrams of each other, and falsy otherwise.


  • Input will only contain letters (ASCII 65 to 90 and 97 to 122), digits (ASCII 48 to 57) or space (ASCII 32).
  • The anagram relation is independendent of case. So "Arm" and "RAM" are anagrams.
  • Spaces don't count either. So "keyboard" and "Barked Yo" are anagrams
  • All builtins allowed
  • Input format is flexible (two strings, an array of two strings, a string containing both phrases with a suitable separator ...)

Code golf. Fewest bytes wins.

Test cases


Lynn, Nyl N
Digital Trauma, Tau Digital Arm
Sp3000, P S 3000
Manage Trash So, Those anagrams


Calvins Hobbies, Helka Homba
Android, rains odd
In between days, bayed entwine
Code golf, cod elf got
  • 8
    \$\begingroup\$ Related but different (only letters, no case, no spaces) \$\endgroup\$
    – Luis Mendo
    Feb 25 '16 at 16:09
  • 4
    \$\begingroup\$ This question's title is very perplexing to someone who's not had enough coffee. +1 :D \$\endgroup\$
    – cat
    Feb 25 '16 at 17:36
  • 1
    \$\begingroup\$ @DonMuesli I would argue that this is still a dupe. The slight changes are very trivial. \$\endgroup\$
    – user45941
    Feb 25 '16 at 17:59
  • 15
    \$\begingroup\$ Manage Trash So, Those anagrams. Nice. \$\endgroup\$
    – mbomb007
    Feb 25 '16 at 20:13
  • 3
    \$\begingroup\$ So, the anagrams... \$\endgroup\$
    – Leaky Nun
    May 21 '16 at 17:23

32 Answers 32


Retina, 25


Try it Online! Additionally, you can run a modified multi-line version.

Delete letters from before the comma along with their matches after the comma. If we have no letters left then it was an anagram.

  • \$\begingroup\$ For Retina, if a positive number could be considered a failure, and zero be considered success, this could be three bytes shorter by using \w as the last stage. \$\endgroup\$ Feb 25 '16 at 20:20
  • \$\begingroup\$ \W will not work for the case of: Calvins Hobbies, Calvin's Hobbies \$\endgroup\$
    – andlrc
    Feb 26 '16 at 0:34
  • \$\begingroup\$ @dev-null "Input will only contain letters (ASCII 65 to 90 and 97 to 122), digits (ASCII 48 to 57) or space (ASCII 32)" \$\endgroup\$ Feb 26 '16 at 2:36

05AB1E, 9 8 bytes




l         # Lowercase the strings
 vy   }   # Map over the list, for each...
   ð-     #   remove spaces
     {    #   and sort
       Q  # Check equality

Try it online!

  • \$\begingroup\$ ... but 2 bytes less! Well done! \$\endgroup\$
    – Luis Mendo
    Feb 25 '16 at 16:28
  • \$\begingroup\$ It's lvyðK{}Q now. \$\endgroup\$ Jun 23 '17 at 16:38

Pyth, 11 10 bytes

Thanks to @FryAmTheEggman for teaching me the power of ;!


Try it here!

Takes a list of two strings as input.


qFmSr-d;0Q    # Q = input

  m      Q    # map Q with d as lambda variable
     -d;      # filter spaces out of the string
    r   0     # convert to lowercase
   S          # sort all characters in string
qF            # Unfold resulting list and check for equality

Python 2, 63 61 bytes

lambda*l:len({`sorted(s.lower())`[2::5].strip()for s in l})<2

An anonymous function that, in fact, takes n arguments and determines if all n of them are mutual palindromes! f("Lynn", "Nyl N") returns True.

This set comprehension trick is by xnor. It saved two bytes, but the old approach looked very neat:

exec"a=`sorted(input().lower())`[2::5].strip();a"*2;print a==aa
  • \$\begingroup\$ `sorted(input().lower())`.strip(" [',") is the same length :/ \$\endgroup\$
    – Sp3000
    Feb 25 '16 at 16:43
  • \$\begingroup\$ The exec thing is clever but seems too complex. You can do better with lambda*l:len({`sorted(s.lower())`[2::5].strip()for s in l})<2. \$\endgroup\$
    – xnor
    Feb 25 '16 at 16:45
  • 2
    \$\begingroup\$ Thanks! I'm a bit disappointed – it looked very cool. Keeping it in the post anyway. \$\endgroup\$
    – Lynn
    Feb 25 '16 at 16:50

Jelly, 12 bytes


Try it online!

How it works

ḟ€⁶O&95Ṣ€QLḂ  Main link. Input: A (list of strings)

  ⁶           Yield ' '.
ḟ€            Filter it from each string.
   O          Apply ordinal to all characters.
    &95       Take bitwise AND with 95 to make the ordinals case-insensitive.
       Ṣ€     Sort each list of ordinals.
         Q    Deduplicate the list.
          L   Get the length.
           Ḃ  Compute the length's parity (1 -> 1, 2 -> 0).

Alternate version, non-competing (9 bytes)

Jelly's uppercase atom had a bug, and Jelly still had no built-in to test lists for equality...


Try it online!

How it works

ḟ⁶ŒuṢµ€⁼/     Main link. Input: A (list of strings)

     µ€       Map the chain to the left over A.
 ⁶            Yield ' '.
ḟ             Filter it from the string.
  Œu          Cast to uppercase.
    Ṣ         Sort.
       ⁼/     Reduce by equality.

CJam, 11 12 14 bytes

3 2 bytes removed thanks to @FryAmTheEggman


Try it online!

{      }2*       e# do this twice
 l               e# read line as a string
  el             e# make lowercase
    S-           e# remove spaces from string
      $          e# sort
          =      e# compare strings
  • \$\begingroup\$ @FryAmTheEggman Thank you! \$\endgroup\$
    – Luis Mendo
    Feb 25 '16 at 16:56
  • \$\begingroup\$ @FryAmTheEggman Thanks again! I still have much to learn about CJam :-) \$\endgroup\$
    – Luis Mendo
    Feb 25 '16 at 16:59
  • 3
    \$\begingroup\$ Your code is secretly laughing. lel. \$\endgroup\$
    – Cyoce
    Feb 27 '16 at 21:12
  • \$\begingroup\$ Or is it a one? lel ==> 1e1 No one knows. It is a mystery. \$\endgroup\$
    – user48538
    Mar 2 '16 at 12:23

Javascript, 69 61 60 59 bytes

1 byte off thanks @ӍѲꝆΛҐӍΛПҒЦꝆ. 1 byte off with currying (pointed out by @apsillers)

n=>m=>(G=s=>[]+s.toLowerCase().split(/ */).sort())(n)==G(m)

        .split(/ */)

F=(n,m)=>document.body.innerHTML+=`<pre>f('${n}')('${m}') -> ${f(n)(m)}</pre>`

F('Luis Mendo','Don Muesli')
F('Calvins Hobbies','Helka Homba')
F('Android','rains odd')
F('In between days','bayed entwine')
F('Code golf','cod elf got')
F('Lynn','Nyl N')
F('Digital Trauma','Tau Digital Arm')
F('Sp3000','P S 3000')
F('Manage Trash So','Those anagrams')

  • 2
    \$\begingroup\$ Very nice, filtering out spaces and converting to an array at the same time! \$\endgroup\$
    – Neil
    Feb 25 '16 at 17:06
  • 2
    \$\begingroup\$ Very nice. You can save one byte using currying, which the community has decided is an acceptable form of arguments: n=>m=>... \$\endgroup\$
    – apsillers
    Feb 25 '16 at 17:23
  • \$\begingroup\$ Try n=>m=>(G=s=>[]+s.toLowerCase().split(/\S/).sort())(n)==G(m). Using split instead of match should save you a byte. \$\endgroup\$ Feb 26 '16 at 3:42
  • \$\begingroup\$ @ӍѲꝆΛҐӍΛПҒЦꝆ. No, because suppose s='db cz'... Now s.match(/\S/g).sort() results in ['b','c','d','z']... and s.split(/\s/).sort() results in ['cz','db'] \$\endgroup\$
    – removed
    Feb 26 '16 at 12:59
  • \$\begingroup\$ @ӍѲꝆΛҐӍΛПҒЦꝆ. But... looking into your idea, I changed it a bit and saved one byte... thanks! \$\endgroup\$
    – removed
    Feb 26 '16 at 13:05

MATL, 11 bytes


EDIT (May 20, 2016) The code in the link uses Xz instead of Xv, owing to recent changes in the language.

Try it online!

2:"     ]       % do this twice
   j            % read input line as a string
    k           % convert to lowercase
     Xv         % remove spaces
       S        % sort
         X=     % are they equal?
  • 3
    \$\begingroup\$ Did you just change your name for that challenge? \$\endgroup\$
    – Denker
    Feb 25 '16 at 16:34
  • 3
    \$\begingroup\$ @DenkerAffe I had been thinking about it for some time. I just made it coincide with the challenge :-) \$\endgroup\$
    – Luis Mendo
    Feb 25 '16 at 16:36
  • 1
    \$\begingroup\$ Don Muesli lol. So you are the Lord of Muesli Luis!? Is this how you keep your healthy complexion? \$\endgroup\$
    – rayryeng
    Feb 25 '16 at 21:00
  • \$\begingroup\$ @rayryeng Heyyy! Good to see you here, Ray! Get back to golfing! \$\endgroup\$
    – Luis Mendo
    Feb 25 '16 at 21:22
  • \$\begingroup\$ I promise I will :) once this course ends... I see you are learning CJam now too. Very nice! \$\endgroup\$
    – rayryeng
    Feb 25 '16 at 21:32

Seriously, 11 9 bytes


Try It Online!

Everyone seems to be using the same algorithm. Here it is yet again.

2`    `n          Do it twice
  ,               Read a string
   ù              Make it lowercase
    S             Sort
     ô            Strip spaces.
        =         Check equality.

Edit: realized sorting does work correctly on strings, and sorts spaces to the front so strip() will work.


C, 165 bytes

#define d(x) int x(char*a,char*b){
#define n(x) for(qsort(x,strlen(x),1,(__compar_fn_t)q);*x<33;x++);
d(s)n(a)n(b)return strcasecmp(a,b);}

Readable and in working context,

#include <stdlib.h>
#include <string.h>
#include <stdio.h>

// start of comparison
int q(char *a, char *b){
     return ((*a)&0xdf)-((*b)&0xdf); // case-insensitive
int s(char *a, char *b){
    for(qsort(a,strlen(a),1,(__compar_fn_t)q); *a<33; a++) /**/;
    for(qsort(b,strlen(b),1,(__compar_fn_t)q); *b<33; b++) /**/;
    return strcasecmp(a,b);
// end of comparison

int main(int i, char **v){
    printf("'%s' '%s'", v[1], v[2]);
    printf("=> %d\n", s(v[1], v[2])); // 0 if equalish
    return 0;

zsh, 85 bytes

[ $(for x in $@;{tr -d \ <<<$x|tr A-Z a-z|fold -1|sort|paste -sd x}|uniq|wc -l) = 1 ]

Input as command line arguments, output as return code.

The for syntax makes this Bash-incompatible.

[               # test...
$(for x in $@;  # map over arguments
{tr -d \ <<<$x  # remove spaces
|tr A-Z a-z     # lowercase
|fold -1        # put each character on its own line
|sort           # sort lines
|paste -sd x    # remove all newlines except last
}|uniq          # take only unique lines
|wc -l          # how many lines remain?
) = 1 ]         # if only 1 line left, it must have been an anagram

Japt, 12 bytes

N®v ¬n ¬xÃä¥

Test it online!

How it works

        // Implicit: N = array of input strings
N®    Ã // Take N, and map each item Z to:
v ¬n    //  Take Z.toLowerCase(), split into chars, and sort.
¬x      //  Join and trim off whitespace.
ä¥      // Reduce each pair of items (that's exactly one pair) X and Y to X == Y.

GNU Sed, 33

Score includes +2 for -rn options to sed.

This is almost a direct port of @FryAmTheEggman's Retina answer:




Perl, 34 33 + 1 = 34 bytes


Requires the -n flag and the free -M5.010|-E:

$ perl -M5.010 -ne's/(.)(.*,.*)\1/$2/i?redo:say!/\w/' <<< 'hello, lloeh'

How it works:

                                   # '-n' make a implicit while loop around the code
 s/(.)(.*,.*)\1/$2/i               # Remove a letter that occurs on both sides of the comma.
                     redo:         # Redo is a glorified goto statement that goes to the top of the while loop
                          say!/\w/ # Check to see if any letter is left

Thanks to msh210 for suggesting using ternary operators to save one byte


Baloch Gyr, 9 bytes


Try it online!

Truthy/falsy output is achieved through predicate success/failure, this being Brachylog.

Previously saved a byte using cṇ₁cḷḍ instead of {ṇ₁cḷ}ᵐ under the assumption that the two input strings would be the same length minus whitespace, but I realized that it would succeed where it should fail on Ah Hass, haha.

{    }ᵐ      For both strings in the input,
 ṇ₁          split on spaces,
   c         concatenate,
    ḷ        and lowercase.
       pᵈ    The strings are permutations of each other.

PHP, 109 94 bytes

function f($x){return str_split((trim($x));}function g($x,$y){return array_diff(f($x),f($y));}

Blech, the two function/returns are killing me here.

Returns the difference between two string inputs as an array of characters. PHP considers [] falsy, satisfying the return requirements.

  • 3
    \$\begingroup\$ function($x,$y){$S=str_split;return array_diff($S(trim($x)),$S(trim($y)));} --> 75 bytes. Creates an anonymous function that returns the result. I've removed that long function and replaced the calls to str_split with an assignmed variable, to shorten it up. \$\endgroup\$ Feb 25 '16 at 16:52
  • \$\begingroup\$ Nice. I was tweaking it to reduce it to the one function, this is two steps ahead of that, well done. \$\endgroup\$
    – ricdesi
    Feb 25 '16 at 16:56

Bash + GNU utilities, 51

f()(fold -1<<<${@^^}|sort)
f $1|diff -qBw - <(f $2)
  • Define a function f() which:
    • ${@^^} converts all parameters to upper case
    • fold -1 splits chars - one per line
    • sorts lines
  • call diff with -q to suppress full diff output and -Bw to ignore whitespace changes

Pyke (commit 30, noncompetitive), 9 bytes



F      )  -  for _ in eval_or_not(input())
 l1       -     ^.lower()
   dk:    -    ^.replace(" ", "")
      S   -   sorted(^)
        q - ^==^

Mathematica, 77 76 bytes

StringMatchQ[##,IgnoreCase->1>0]&@@(""<>Sort[Characters@#/." "->""]&/@{##})&

The first part is actually one of my answers to another question!


Pike, 54 112 109 109 96 bytes

#define a(x) sort((array)replace(lower_case(x)," ",""))
int s(mixed i){return a(i[0])==a(i[1]);}

mixed happens to be shorter than array(string).

s returns 1 if its arguments are anagrams.


APL, 31 chars

{≡/{x[⍋x←('.'⎕R'\u0')⍵~' ']}¨⍵}

To be used so:

    {≡/{x[⍋x←('.'⎕R'\u0')⍵~' ']}¨⍵}'Sp3000' 'P S 3000' 

In English:

  • { ... }¨⍵: for each of the two elements of the argument
  • x←('.'⎕R'\u0')⍵~' ': transform to uppercase (using a regex...) the string without the spaces and assign the temporary result to x
  • x[⍋x]: sort x
  • ≡/: compare the two results of the sorting: if they match, return 1.
  • \$\begingroup\$ Is it possible to try it online? I tried with this but I don't really know how to use it \$\endgroup\$
    – Luis Mendo
    Jun 28 '16 at 11:49
  • \$\begingroup\$ Sure. Here: definition after which you can just type f 'first avatar' 'second avatar' \$\endgroup\$
    – lstefano
    Jun 28 '16 at 12:56
  • \$\begingroup\$ Thanks! Maybe add that to the answer? So that people can try \$\endgroup\$
    – Luis Mendo
    Jun 28 '16 at 13:28
  • \$\begingroup\$ –9: ≡/{x[⍋x←0~⍨32|⎕UCS⍵]}¨ \$\endgroup\$
    – Adám
    Jun 28 '16 at 15:27
  • \$\begingroup\$ @Adám: that won't work because ≡/{x[⍋x←0~⍨32|⎕UCS⍵]}¨'pp' '00' gives 1. \$\endgroup\$
    – lstefano
    Jun 29 '16 at 8:52

Java, 218 Bytes

First time I've ever written Java...


import java.util.Arrays;boolean M(String a,String b){char[]A=a.toUpperCase().replace(" ","").toCharArray();char[]B=b.toUpperCase().replace(" ","").toCharArray();Arrays.sort(A);Arrays.sort(B);return Arrays.equals(A,B);}


import java.util.Arrays;
public class ManageTrashSo {
    public boolean M(String a, String b) {
    char[] A = a.toUpperCase().replace(" ", "").toCharArray();
    char[] B = b.toUpperCase().replace(" ", "").toCharArray();
    return Arrays.equals(A, B);


    ManageTrashSo manageTrashSo = new ManageTrashSo();

    System.out.println(manageTrashSo.M("Lynn", "Nyl N"));
    System.out.println(manageTrashSo.M("Digital Trauma", "Tau Digital Arm"));
    System.out.println(manageTrashSo.M("Android", "rains odd"));
    System.out.println(manageTrashSo.M("In between days", "bayed entwine"));
  • \$\begingroup\$ I know it's been almost a year, but you can golf it by 32 bytes like this: boolean f(String...a){java.util.Arrays x=null;String[]A=g(a[0]),B=g(a[1]);x.sort(A);x.sort(B);return x.equals(A,B);}String[]g(String a){return a.replace(" ","").toUpperCase().split("");} (186 bytes) Or if you convert it to a Java 8 lambda, it can be: a->b->{java.util.Arrays x=null;String[]A=g(a),B=g(b);x.sort(A);x.sort(B);return x.equals(A,B);};String[]g(String a){return a.replace(" ","").toUpperCase().split("");} (167 bytes). Here is a TIO with test code. \$\endgroup\$ Sep 27 '17 at 11:48

Japt, 10 bytes

v á øVrS v

Try it


Ruby, 50 bytes

def f;gets.upcase.chars.sort.join.strip;end
p f==f

Writing f=->{...} and f[]==f[] is just as long. :(


PowerShell, 81 bytes

param([char[]]$a,[char[]]$b)-join($a-replace' '|sort)-eq-join($b-replace' '|sort)

A slight rewrite of my answer on the linked Anagram challenge.

Takes input as char-arrays, performs a -replace operation to remove spaces, sorts them (which sorts alphabetically, not by ASCII value), then -joins them back into a string. The -eq in PowerShell is by default case-insensitive, but here it must be performed on strings, as [char]'a' is not equal to [char]'A', hence the reason for -join.


Perl, 35 bytes

Include +1 for -p

Somewhat abusive since it depends on the program being given on the commandline.

perl -pe'<>=~s%\S%*_=s/$&//i?_:0%reg;$_=!//'

Then give the strings as 2 consecutive lines on STDIN

A very abusive solution is 30 bytes:

perl -ne'<>=~s%\w%1/!s/$&//i%reg;1/!//'

This crashes if the strings are not anagrams and therefore gives a false exit code from the point of view of the shell. It also gives garbage on STDERR for that case. If the strings are anagrams the program is silent and gives a "true" exit code


Q, 25 Bytes


NOTE.- counting include function name f: to facilitate tests (as lambda we can decrement 2 Bytes)

Readable version

match over {ascending not null lower x} each x

{.. x ..} is an anonymous function with arg x
_x        lowers string x
&~^x      where not null x (space is considered null)
x@..      selects elements of x according to indexes .. 
<x        ascending indexes of x (not values). Ex <"cab" is 1 2 0
x@<x      ascending values of x (x at ascending indexes of x)
~         match (diad function). Ex "one"~"one" is true
f'..      applies function f for each argument ..
f/..      applies function f over elements of sequence (fold)


f("Lynn";"Nyl N")                       
f("Digital Trauma";"Tau Digital Arm")   
f("Sp3000";"P S 3000")                  
f("Manage Trash So";"Those anagrams")   
f("Calvins Hobbies";"Helka Homba")      
f("Android";"rains odd")                
f("In between days";"bayed entwine")    
f("Code golf";"cod elf got")    

generates (1b = true, 0b = false)


About Q

General-purpose language (APL derivative, specialized in data processing) developed by kx.com. Free full functional evaluation version for Windows/Linux/MacOS.

  • \$\begingroup\$ What do you mean, other languages are not serious? :-P \$\endgroup\$
    – Luis Mendo
    May 21 '16 at 8:59
  • \$\begingroup\$ If the f is required for the code to evaluate properly, then it must be counted. Otherwise, just leave it off of your submission code, and only use it in examples to show how to assign the function. \$\endgroup\$
    – user45941
    May 21 '16 at 9:28
  • \$\begingroup\$ Of course, other languages are as serious as Q. I beg my poor english. But some languages sacrifices readability or are equiped with libraries ad-hoc for this type of contests. Q is a 'general purpose language', in spite of the fact that code is not very readable. \$\endgroup\$
    – J. Sendra
    May 21 '16 at 16:59
  • \$\begingroup\$ You only need to assign x once if you lower later, thus k)~/{x@<x:_x@&~^x}' for 17 bytes.. but I'd say it's 19 as you need the k) bracket as this is K code rather than Q... \$\endgroup\$
    – mkst
    Sep 27 '17 at 14:15

PHP, 89 bytes

for(;$i++<2;)$r[]=count_chars(join(explode(" ",strtolower($argv[$i]))));echo$r[0]==$r[1];

Try it online!

PHP, 94 bytes


Try it online!


Excel VBA, 122 Bytes

Anonymous VBE immediate window Function that takes input from range [A1:B1] and outputs to the VBE immediate window

a=Replace([A1]," ",""):b=Replace([B1]," ",""):For i=1To Len(a):b=Replace(b,Mid(a,i,1),"|",,1,1):Next:?b=String(len(a),"|")

C#, 378 bytes

I need a handicap!!


using System;
using System.Linq;
using System.Text;

public class Program

    public static void Main()
        var l = "Hello World";

        var r = "Red Who Loll";

        var y = new Func<string,string>(s => new String(s.ToLower().Replace(" ","").OrderBy(v => v).ToArray()));
        var z = new Func<string,string,Func<string,string>,bool>((w,x,f) => f(w) == f(x));
        var o = z(l, r, y);

        Console.WriteLine("{0} & {1} are anagram: {2}",l, r, o);

                Console.WriteLine("C#, {0} bytes", Encoding.Unicode.GetByteCount(@"var y = new Func<string,string>(s => new String(s.ToLower().Replace("" "","""").OrderBy(v => v).ToArray()));
    var z = new Func<string,string,Func<string,string>,bool>((w,x,f) => f(w) == f(x));"));


  • 4
    \$\begingroup\$ Welcome to Programming Puzzles & Code Golf! General rule is to put your language together with your byte count in the headline of your post. You can do that by adding a leading # to the first line. Also for code-golf questions it is required to golf your program. For the start you should remove unecessary whitespaces and use single-character variable names. Also you can always use a function instead of a full program (unless explicitly forbidden) to save some more bytes. \$\endgroup\$
    – Denker
    Feb 28 '16 at 0:40
  • \$\begingroup\$ @DenkerAffe you ninja'd me :) \$\endgroup\$
    – cat
    Feb 28 '16 at 0:40
  • 1
    \$\begingroup\$ Why all the whitespace?? \$\endgroup\$ Mar 6 '16 at 20:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.