24
\$\begingroup\$

The title of Numberphile's newest video, 13532385396179, is a fixed point of the following function f on the positive integers:

Let n be a positive integer. Write the prime factorization in the usual way, e.g. 60 = 22 · 3 · 5, in which the primes are written in increasing order, and exponents of 1 are omitted. Then bring exponents down to the line and omit all multiplication signs, obtaining a number f (n). [...] for example, f (60) = f (22 · 3 · 5) = 2235.

(The above definition is taken from Problem 5 of Five $1,000 Problems - John H. Conway)

Note that f (13532385396179) = f (13 · 532 · 3853 · 96179) = 13532385396179.

Task

Take a positive composite integer n as input, and output f(n).

Another example

48 = 24 · 3, so f (48) = 243.

Testcases

More testcases are available here.

   4 -> 22
   6 -> 23
   8 -> 23
  48 -> 243
  52 -> 2213
  60 -> 2235
 999 -> 3337
9999 -> 3211101
\$\endgroup\$
  • 11
    \$\begingroup\$ +1 I'm still astonished that someone managed to find 13532385396179 as a disproof of the conjecture. I guess the $1000 prize would go some way to pay for the electricity used! :) \$\endgroup\$ – Wossname Jun 9 '17 at 10:21
  • 7
    \$\begingroup\$ Without following the link it wasn't clear that the conjecture is that repeated applications of f(n) will always reach a prime (and of course f(p) = p if p is prime). 13532385396179 disproves the conjecture because it's both composite and a fixed opint. \$\endgroup\$ – Chris H Jun 9 '17 at 12:59

20 Answers 20

16
\$\begingroup\$

Python, 166 162 159 bytes

You guys are much better. This is what I used! (the algorithm that solved it calls this)

from primefac import*
def c(n):
 x=factorint(n)
 a=''
 for i in range(len(x)):
  l=min(x.keys())
  a+=str(l)
  if x[l]>1:a+=str(x[l])
  x.pop(l)
 return int(a)
\$\endgroup\$
  • 2
    \$\begingroup\$ Why did someone downvote a newcomer instead of helping him improve his answer as @LeakyNun did? :( \$\endgroup\$ – Shaggy Jun 9 '17 at 11:30
  • 3
    \$\begingroup\$ Sorry- that really is what I used (I found the number). I just thought the crummy code would be funny. You can take it down. \$\endgroup\$ – jchd Jun 9 '17 at 11:33
  • 9
    \$\begingroup\$ Welcome on the site. It's really nice to have you sharing with us your solution. (for people who don't know, Jim Davis is the one who solved this problem in the first place). However, answers to challenges need to follow some rules. If you just follow the suggestions from @LeakyNun, then you answer will be valid. (maybe have a look at the other answers to see how they usually look like) \$\endgroup\$ – Dada Jun 9 '17 at 11:47
  • 4
    \$\begingroup\$ Oh my God, I didn't expect Jim Davis himself to appear in this site, and to answer my challenge... I feel so honoured now... \$\endgroup\$ – Leaky Nun Jun 9 '17 at 13:15
  • 2
    \$\begingroup\$ ehh, not a troll by the way. My email address is on gladhoboexpress.blogspot.ca/2014/10/climb-to-prime.html ... I left the post up, nobody swamps you with email over math. \$\endgroup\$ – jchd Jun 9 '17 at 13:38
9
\$\begingroup\$

Brachylog, 8 bytes

ḋoọc;1xc

Try it online!

Explanation

Example input: 60

ḋ          Prime decomposition: [5,3,2,2]
 o         Order: [2,2,3,5]
  ọ        Occurences: [[2,2],[3,1],[5,1]]
   c       Concatenate: [2,2,3,1,5,1]
    ;1x    Execute 1s: [2,2,3,5]
       c   Concatenate: 2235

You can use ℕ₂ˢ (select all integers greater than or equal to 2) instead of ;1x, which is probably more readable and more in the spirit of Brachylog.

\$\endgroup\$
9
\$\begingroup\$

Jelly, 6 bytes

ÆFFḟ1V

Try it online!

Explanation

ÆF      Get prime factorisation of input as prime-exponent pairs.
  F     Flatten.
   ḟ1   Remove 1s.
     V  Effectively flattens the list into a single integer.
\$\endgroup\$
  • \$\begingroup\$ V = "concatenate to a single string and eval as Jelly" \$\endgroup\$ – Erik the Outgolfer Jun 9 '17 at 10:48
  • \$\begingroup\$ @EriktheOutgolfer Yes, hence "effectively". \$\endgroup\$ – Martin Ender Jun 9 '17 at 10:49
  • \$\begingroup\$ @MartinEnder Any particular reason you don't use (Convert from decimal to integer)? \$\endgroup\$ – scatter Jul 3 '17 at 19:00
  • \$\begingroup\$ @Christian Because the list might contain multi-digit integers. \$\endgroup\$ – Martin Ender Jul 3 '17 at 19:26
  • \$\begingroup\$ @MartinEnder Ah, clever. I've used FḌ in the past - that's a good tip! \$\endgroup\$ – scatter Jul 4 '17 at 0:18
5
\$\begingroup\$

Mathematica, 43 36 Bytes

Row@Flatten@FactorInteger@#/. 1->""&

Try it online!

\$\endgroup\$
  • 2
    \$\begingroup\$ DeleteCases is long, you can use /.1->"" or /.1->##&[] (alternative form of /.1->Nothing \$\endgroup\$ – user202729 Jun 9 '17 at 10:42
  • 3
    \$\begingroup\$ @user202729 All of those need a space in front of the 1 to prevent it from parsing as ... / (0.1). \$\endgroup\$ – Martin Ender Jun 9 '17 at 10:48
  • \$\begingroup\$ You are right! fixed \$\endgroup\$ – J42161217 Jun 9 '17 at 14:40
4
\$\begingroup\$

CJam, 8 bytes

limF:~1-

Try it online!

Explanation

li  e# Read input and convert to integer.
mF  e# Get prime factorisation as prime-exponent pairs.
:~  e# Flatten.
1-  e# Remove 1s.
    e# Implicitly print a flattened representation of the list.
\$\endgroup\$
  • \$\begingroup\$ I would have used e_ to flatten, since that's what it's there for, but it doesn't change the score. \$\endgroup\$ – Peter Taylor Jun 9 '17 at 14:31
  • 1
    \$\begingroup\$ @PeterTaylor Hm yeah, I can never decide which one to use, but tend to go with e_ for deep flatten only and use :~ whenever it's just a single level. \$\endgroup\$ – Martin Ender Jun 9 '17 at 14:36
4
\$\begingroup\$

05AB1E, 10 bytes

Òγʒ¬?gDië?

Try it online!

Ò          # Push list of prime factors with duplicates
 γ         # Break into chunks of consecutive elements
  ʒ        # For each
   ¬?      #   Print the first element
     gD    #   Push the length of this chunk twice
       ië  #   If not 1
         ? #     Print the length
\$\endgroup\$
3
\$\begingroup\$

05AB1E, 12 11 bytes

Òγvy¬sgD≠×J

Try it online!

Explanation

Ò            # calculate prime factors with duplicates
 γ           # group consecutive equal elements
  vy         # for each group
    ¬        # get the head without popping
     sg      # push the length of the group
       D≠×   # repeat the length (length != 1) times
          J  # join
\$\endgroup\$
  • \$\begingroup\$ Fails for 48. \$\endgroup\$ – Leaky Nun Jun 9 '17 at 10:02
2
\$\begingroup\$

Pyth, 12 bytes

smjk_>hddr8P

Try it!

alternative, 12 bytes

smjk<_AdGr8P

Try that!

explanation

smjk_>hddr8P
           PQ  # prime factorization (already in correct order) of the implicit input: [3, 3, 11, 101]
         r8    # length encode: [[2, 3], [1, 11], [1, 101]]
 m             # map over the length encoded list (lambda variable: d)
     >hdd      # take the d[0] last elements of d (so only the last for d[0]==1 and all else)
    _          # reverse that list
  jk           # join into a string
s              # conatenate the list of strings
\$\endgroup\$
2
\$\begingroup\$

Pyth, 11 bytes

jksm_-d1r8P

Try here

\$\endgroup\$
2
\$\begingroup\$

Python 2, 99 bytes

n=input()
r=''
p=2
while~-n:
 e=0
 while n%p<1:e+=1;n/=p
 r+=str(p)*(e>0)+str(e)*(e>1);p+=1
print r

Try it online!

If inputs are restricted to be below 2147483659, both str(...) may be replaced by `...` saving 6 bytes (this program will be very slow for numbers affected anyway!).

\$\endgroup\$
2
\$\begingroup\$

Ohm, 11 bytes

o:_]D2<?O;J

Try it online!

Explanation

o:_]D2<?O;J
o           # Push prime factors with powers from input (Format [[prime,power],...]
 :          # For each...
  _          # Push current element
   ]         # flatten
    D        # Duplicate power
     2<? ;   # Is the power smaller than 2?
        O     # Delete top of stacks
          J  # Join
\$\endgroup\$
1
\$\begingroup\$

Japt, 19 bytes

k ó¥ ®¯1 pZlÃc fÉ q

Test it online!

Explanation

 k ó¥  ®   ¯  1 pZlà c fÉ  q
Uk ó== mZ{Zs0,1 pZl} c f-1 q  // Ungolfed
                              // Implicit: U = input number
Uk                            // Break U into its prime factors.
   ó==                        // Group into runs of equal items.
       mZ{         }          // Map each item Z in this to
          Zs0,1               //   Z.slice(0, 1) (the array of the first item),
                pZl           //   with Z.length added at the end.
                              // This returns an array of prime-exponent pairs (Jelly's ÆF).
                     c        // Flatten.
                       f-1    // Filter to the items X where X - 1 is truthy (removes '1's).
                           q  // Join the resulting array into a single string.
                              // Implicit: output result of last expression
\$\endgroup\$
1
\$\begingroup\$

PHP, 88 bytes

for($i=2;1<$a=&$argn;)$a%$i?$i++:$a/=$i+!++$r[$i];foreach($r as$k=>$v)echo$k,$v<2?"":$v;

Try it online!

\$\endgroup\$
0
\$\begingroup\$

C#, 206 100 bytes

n=>{var r="";for(int d=1,c;++d<=n;){c=0;while(n%d<1){c++;n/=d;}r+=c>0?d+(c>1?c+"":""):"";}return r;}

Full/Formatted version:

using System;

class P
{
    static void Main()
    {
        Func<int, string> func = n =>
        {
            var r = "";
            for (int d = 1, c; ++d <= n;)
            {
                c = 0;
                while (n % d < 1)
                {
                    c++;
                    n /= d;
                }

                r += c > 0 ? d + (c > 1 ? c + "" : "") : "";
            }

            return r;
        };

        Console.WriteLine(func(4));
        Console.WriteLine(func(6));
        Console.WriteLine(func(8));
        Console.WriteLine(func(48));
        Console.WriteLine(func(52));
        Console.WriteLine(func(60));
        Console.WriteLine(func(999));
        Console.WriteLine(func(9999));

        Console.ReadLine();
    }
}
\$\endgroup\$
0
\$\begingroup\$

Javascript - 91 bytes

(x,r='',i=1,n)=>{while(x>i++){for(n=0;x%i<1;n++)x/=i;r+=(n>0?i+'':'')+(n>1?n:'')}return r}

Explanation

(x,r='',i=1,n)=>(          // input x is the number to process, r, i, n are default values only
    while(x>i++){          // iterate over i until x
        for(n=0;x%i<1;n++) // iterate over n until i is not a factor of x
            x/=i;          // factor i out of x
        r+=(n>0?i+'':'')   // append i to r if n > 0
            +(n>1?n:'')    // append n to r if n > 1
                           // i+'' prevents adding i and n before appending to r
    }
    return r               // return r by comma-operator and arrow function syntax
)
\$\endgroup\$
0
\$\begingroup\$

Java 8, 103 chars

Pretty straightforward solution.

n->{String r="";int d=2,c;while(n>1){c=0;while(n%d<1){c++;n/=d;}if(c>0)r+=d;if(c>1)r+=c;d++;}return r;}

Ungolfed:

private static Function<Integer, String> f = n->{
    String result = "";
    int divisor = 2, count;
    while (n>1) {
        count = 0;
        while (n % divisor < 1) {
            count++;
            n /= divisor;
        }
        if (count > 0) result += divisor;
        if (count > 1) result += count;
        divisor++;
    }
    return result;
};
\$\endgroup\$
0
\$\begingroup\$

Octave, 69 bytes

@(a)printf('%d',(f=[[~,c]=hist(b=factor(a),d=unique(b));d](:))(f~=1))

Try it online!

Ended up being quite long, but this will generate the desired output.

Essentially we use the histogram function to count the number of occurrences of the unique values in the prime factorisation of the input value.

  • The result of the factor() function gives the prime factors in ascending order
  • we then find unique() values in that array
  • hist() returns the number of occurrences

Once we have the two arrays (one for unique factors, one for counts), we concatenate the arrays vertically (one on top of the other), and then flatten. This interleaves the factors with counts.

Finally we display the result as a string ensuring to skip any 1's in the final array. The only time 1's can appear is if the count was 1 because 1 will never be a prime factor. This elimination is done before converting to a string so it won't affect things like the number 10.

\$\endgroup\$
0
\$\begingroup\$

Ruby, 45 + 7 bytes

Requires the flag -rprime.

->n{(n.prime_division.flatten-[1]).join.to_i}

Try it online!

\$\endgroup\$
0
\$\begingroup\$

Pyth - 16 bytes

V{PQpNK/PQNItKpK

Try it

Another solution:

sm`d-.nm(d/PQd){PQ1
\$\endgroup\$
  • 1
    \$\begingroup\$ One can replace FN by V. \$\endgroup\$ – Leaky Nun Jun 11 '17 at 17:34
  • 1
    \$\begingroup\$ Also, r8 (run-length encoding) seems to be useful. \$\endgroup\$ – Leaky Nun Jun 11 '17 at 17:35
0
\$\begingroup\$

R, 72 bytes

x=rle(pracma::factors(scan()));x$l[x$l<2]='';paste0(x$v,x$l,collapse='')

Requires the pracma package, which is not installed on TIO.

\$\endgroup\$

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.