Sort digits by their first occurrence in pi

Given a non-negative number n, sort the digits of n by their first occurrence in pi.

Input can be taken via function cli argument, or STDIN and as a string, char[] or integer. You may output via return value, exit status or STDOUT.

• Related – Stewie Griffin Apr 29 '17 at 16:38
• Can we take input and output as strings, or as arrays of digits? – ETHproductions Apr 29 '17 at 16:39
• @ETHproductions Clarified. – Roman Gräf Apr 29 '17 at 16:40
• A few test cases would be nice. – Dennis Apr 29 '17 at 18:15
• Now that 12 answers are already present, all of which performing the same thing, if you are still unclear what is being asked, then it is not the problem of the question. – Leaky Nun Apr 30 '17 at 2:30

Pyth, 8 6 bytes

ox+.n0


Try it here.

-1 thanks to Leaky Nun: The input will provide the 0 if it's ever needed.
Trivial -1 thanks to Jakube: Backtick not needed (ah, how did I miss that, HOW?!?).

• Woohoo, this even beats 05AB1E! Edit: it doesn't beat 05AB1E, and I don't wanna steal :( – Erik the Outgolfer Apr 29 '17 at 18:00
• I found it. You don't need the 0 at the end. If the input has a 0, the 0 would be provided by the input; if the input does not have a 0, it won't matter. – Leaky Nun Apr 29 '17 at 18:16
• @LeakyNun and you can even save the backtick: ox+.n0 – Jakube Apr 29 '17 at 22:14
• OK, disregard the first comment, thanks to LeakyNun and Jakube, I again beat 05AB1E, I hope for good this time. – Erik the Outgolfer Apr 30 '17 at 8:06
• That's a beautiful amount of implicit input. – isaacg May 13 '17 at 6:27

Python 3, 40 39 bytes

1 byte thanks to Jonathan Allan.

lambda s:sorted(s,key="145926870".find)


Try it online!

• You can drop the 3, since find will return -1 when an item is not found. – Jonathan Allan Apr 29 '17 at 17:51

05AB1E, 109 7 bytes

Saved 1 byte thanks to Leaky Nun noting that filtering out duplicates is unnecessary.
Saved 2 bytes thanks to Adnan.

žqRvy†J


Try it online!

Explanation

žq       # push pi to 15 decimals (contains all digits but 0)
R      # reverse
vy    # for each char in pi
†J  # move it's occurrences in the input to the front

• 13žsRvy†J for 9 bytes – Leaky Nun Apr 29 '17 at 16:29
• @LeakyNun: Oh yeah, duplicates doesn't matter. Thanks :) – Emigna Apr 29 '17 at 16:31
• Can you use žq instead of 13žs? – Adnan Apr 29 '17 at 17:56
• @Adnan It doesn't seem to work. – Erik the Outgolfer Apr 29 '17 at 18:01
• @Adnan: Yes of course. I didn't realize there was another pi constant :) – Emigna Apr 29 '17 at 18:05

Jelly, 10 bytes

“ṀSṪw’ṾiµÞ


Try it online!

Takes input as a string of digits.

-3 bytes thanks to @ETHproductions

Explanation

“ṀSṪw’ṾiµÞ
µ  - Separate chain into function “ṀSṪw’Ṿi and sort atom Þ.
Þ - Sort the input by
i   - Each digit's index in:
“ṀSṪw’     - the literal 3145926870 ...
Ṿ    - transformed into the list 3,1,4,5,9,2,6,8,7,0

• I think 3145926870 can be represented as a 4-digit base-250 string (meaning it takes up 6 bytes instead of 10), but I'm not sure how to compress it as such. – ETHproductions Apr 29 '17 at 16:46
• Does Jelly not have a builtin for pi? – math junkie Apr 29 '17 at 16:46
• @mathjunkie but Jelly is not very efficient on string manipulation – Leaky Nun Apr 29 '17 at 16:47
• @mathjunkie Yes, but the manipulations to the list take too many bytes – fireflame241 Apr 29 '17 at 16:47
• “ṀSṪw’ will give you 3145926870. – Leaky Nun Apr 29 '17 at 16:49

Japt, 10 9 bytes

8 bytes of code, +1 for the -P flag.

¬ñ!bMP+U


Try it online! Takes input as a string.

Explanation

¬ñ!bMP+'0  // Implicit input

¬          // Split the input into chars.
ñ         // Sort each char in the resulting list by
!b       //   its index in
MP+U   //     Math.PI + the input.
-P         // Join the result back into a single string.
// Implicit: output result of last expression


JavaScript (ES6), 54 bytes

f=
s=>[...s].sort((a,b)=>k[a]-k[b],k=9150236874).join
<input oninput=o.textContent=f(this.value)><pre id=o>

Uses strings for I/O.

Jelly,  8  7 bytes

-1 byte thanks to Dennis (use any existing 0 in the input, clever.)

ØP;ṾiµÞ


Try it online!

How?

ØP;ṾiµÞ - Main link: string s (char list)
µÞ - sort the characters, c, of s by:
i   -   first index of c in:
ØP      -     pi yield: 3.141592653589793
;     -     concatenate with left: [3.141592653589793, c]
Ṿ    -     un-evaluate: "3.141592653589793,c" (a char list with the digit character c)
if any c is 0 ^ it will then be to the right of all others

• ...and there I was searching for squares - 3820009 (sqrt of 14592468760081) is still 3 digits in base 250. – Jonathan Allan Apr 29 '17 at 18:25
• The Ṿ in your explanation is misplaced. – Erik the Outgolfer Apr 30 '17 at 9:56
• @EriktheOutgolfer - thanks, adjusted. – Jonathan Allan Apr 30 '17 at 10:38

CJam, 151210 8 bytes

r{P#c}$ Try it online! -3: Use a string based on the P pi variable instead of a literal. -2: Decided I don't need to uniquify at all, since finding an index takes the first occurrence anyways. -2: Thanks to jimmy23013 for an interesting approach using x mod 65536. Explanation: r{P#c}$ e# Takes an input token
r        e# Take the integer as a string
{P#c}  e# Sorting key:
P      e#  Push P (defaults to 3.141592653589793)
     e#  Convert to string representation
#    e#  Find char's index in the string we made
e#  A '.' will never be found in an integer, but it doesn't matter, since the shifting preserves ideal sorting.
e#  A '0' will be indexed as -1.
c   e#  Convert index to char
e#  This first calculates index % 65536, and then converts to char. We need this because otherwise 0 would be indexed as -1, i.e. smallest index.
e#  We don't need to convert back to integer, since we can use lexicographical sorting.
$e# Sort with key PHP, 71 Bytes The regex solution is shorter for(;~$c=_3145926870[$i++];)echo str_repeat($c,substr_count($argn,$c));


or

for(;~$c=_3145926870[$i++];)echo str_pad("",substr_count($argn,$c),$c);  Online Versions PHP, 78 Bytes for(;~$c=$argn[$i++];)$j[strpos("3145926870",$c)].=$c;ksort($j);echo join($j);  PHP, 112 Bytes $a=str_split($argn);usort($a,function($x,$y){return strpos($d="3145926870",$x)<=>strpos($d,$y);});echo join($a);  Online Version • I've added a 69 byte solution. Maybe we can get it down to 66 byte together ;) – Christoph May 2 '17 at 14:13 C, 103 97 bytes char*p="3145926870";s(*a,*b){return strchr(p,*a)-strchr(p,*b);}f(char*t){qsort(t,strlen(t),1,s);}  Try it online • Thanks, @ceilingcat, MSVC does not like this at all, I suppose I should rather prototype with gcc :-) – Johan du Toit May 2 '17 at 18:56 • MSVC probably won't like the fact that gcc lets you ditch char in char*p and char*t – ceilingcat Sep 22 '17 at 20:22 Ruby, 50 bytes n=gets;"3145926870".each_char{|c|$><<c*n.count(c)}


MATL, 14 bytes

YP99Y$uj!y=sY"  Try it online! Explanation with an example The symbol ; is used as row separator in matrices. So [1 2 3] is a row vector, [1; 2; 3] is a column vector, and [1 2; 3 4] is a square matrix. The latter can also be represented, for clarity, as [1 2; 3 4]  Consider input 2325 as an example. YP % Push approximation of pi as a double (predefined literal) % 3.14159265358979 99Y$   % Variable-precision arithmetic with 99 digits. Gives a string.
% The input 3.14159265358979 is recognized as representing pi
% STACK: '3.141592653589793238462 ··· 707'
u      % Unique entries, keeping order of their first appearance
% STACK: '3.145926870'
j      % Input line as a string
% STACK: '3.145926870', '2352'
!      % Transpose
% STACK: '3.145926870', ['2'; '3';'5'; '2']
y      % Duplicate the second-top element in the stack
% STACK: '3.145926870', ['2'; '3';'5'; '2'], '3.145926870'
=      % Test for equality, with broadcast. This gives a matrix with
% all pairwise comparisons)
% STACK: '3.145926870', [0 0 0 0 0 0 1 0 0 0 0;
%                        1 0 0 0 0 0 0 0 0 0 0;
%                        0 0 0 0 1 0 0 0 0 0 0;
%                        0 0 0 0 0 0 1 0 0 0 0]
s      % Sum of each column
% STACK: '3.145926870', [1 0 0 0 1 0 2 0 0 0 0]
Y"     % Run-length decoding. Implicitly display
% STACK: '3522'


C (gcc), 78 bytes

f(char*s){for(char*d="3145926870",*p;*d;d++)for(p=s;*p;)*p++-*d||putchar(*d);}


Try it online!

C# Interactive, 37 36 Bytes

i.OrderBy(c=>"145926870".IndexOf(c))


Actually you have to execute this in the C# interactive for proper results, but I guess this is what you meant with exit status. The variable i actually is the input variable (it can be for example a string), so it's basically the method parameter.

I think the code itself is pretty straight forward.

• Where's the 3? – Paul May 1 '17 at 13:44
• @Paul it's not neccessary, as it returns -1 if the element isn't found. – MetaColon May 1 '17 at 13:47
• This is just a snippet of code though, I'm pretty sure even in interactive you have to specify why i is somewhere so it can be taken as input. Also if you're saying C# you have to include using System.Linq; into the byte count. However, if this is Interactive you should specify the language as C# Interactive not solely C#. – TheLethalCoder May 18 '17 at 13:13
• @TheLethalCoder I updated it to C# Interactive. The using isn't neccesaary in the interactive, as it's included automatically. – MetaColon May 18 '17 at 13:16

05AB1E, 5 6 bytes (noncompeting)

Had to realise that 0 is not present in the standard length pi constant.

Σтžsyk


Try it online!

Σтžsyk
Σ      Sort by the result of code
тžs   Push 100 digits of pi
yk  Index of digit in pi

• You should mark this non-competing as Σ is newer than the challenge. – Emigna May 18 '17 at 13:17
• @Emigna marked it, thanks. But after the required fix it's not anymore shorter than the winning answer anyway ): – kalsowerus May 18 '17 at 13:18
• Too bad you needed that zero for this method. It should be optimal for this language at least. Can't ask for more than that :) – Emigna May 18 '17 at 13:24

PHP, 66 65 bytes

Saved 1 byte thanks to Titus.

while(~$d=_3145926870[++$i])echo preg_filter("/[^$d]/",'',$argn);


Java 7, 110 bytes

String c(String s){String r="";for(char i:"3145926870".toCharArray())r+=s.replaceAll("[^"+i+"]","");return r;}


Explanation:

String c(String s){                       // Method with String parameter and String return-type
String r="";                            //  Result String
for(char i:"3145926870".toCharArray())  //  Loop over the characters of "3145926870"
r+=s.replaceAll("[^"+i+"]","");       //   Append the result-String with all the occurrences of the current character
//  End of loop (implicit / single-line body)
return r;                               //  Return the result-String
}                                         // End of method


Test code:

Try it here.

class M{
static String c(String s){String r="";for(char i:"3145926870".toCharArray())r+=s.replaceAll("[^"+i+"]","");return r;}

public static void main(String[] a){
System.out.println(c("12345678908395817288391"));
}
}


Output:

33311145599922688888770


Clojure, 38 bytes

#(sort-by(zipmap"3145926870"(range))%)


Input in string, returns a sequence of characters. zipmap creates a "dictionary" object, which can be used in a function context as well.

(f "1234")
(\3 \1 \4 \2)


If input digits were guaranteed to be unique then you could simply do #(filter(set %)"3145926870").

PHP, 69 68

for(;(~$d=$argn[$j++])||~$c=_3145926870[$i+++$j=0];)$c==$d&&print$d;  Still beaten by preg_filter but I thought it was quite nice itself. Maybe someone can golf off some bytes. • $c!=$d?:print$d as alternative for $c==$d&&print$d I only see in the moment – Jörg Hülsermann May 2 '17 at 14:21 • _3145926870 instead of "3145926870" save 1 Byte – Jörg Hülsermann May 2 '17 at 14:26 • for(;(~$d=$argn[$j++])?:~$c=_3145926870[++$i+$j=0];$c!=$d?:print$d); is also a working alternative – Jörg Hülsermann May 2 '17 at 14:57

Perl 6, 34 bytes

*.comb.sort:{3145926870.index: $_}  Try it *\ # WhateverCode lambda (this is the parameter) .comb # split into digits .sort: { # sort by 3145926870.index:$_ # its index in this number
}


k, 19 bytes

{x@<"3145926870"?x}


Explanation:

{                 } /function(x)
"3145926870"?x  /for each x: "3145926870".index(x)
<                /get indices with which to sort
x@                 /sort x by those indices
`