Problem 38: Pandigital multiples

(see projecteuler.net/problem=38)

Take the number 192 and multiply it by each of 1, 2, and 3:

192 * 1 = 192
192 * 2 = 384
192 * 3 = 576

By concatenating each product we get the 1 to 9 pandigital, 192384576. We will call 192384576 the concatenated product of 192 and (1,2,3)
The same can be achieved by starting with 9 and multiplying by 1, 2, 3, 4, and 5, giving the pandigital, 918273645, which is the concatenated product of 9 and (1,2,3,4,5).

What is the largest 1 to 9 pandigital 9-digit number that can be formed as the concatenated product of an integer with (1,2, ... , n) where n > 1?

Algorithm

A bitmask tracks whether a number is pandigital (or not): if the n-th bit of bitsUsed is 1 then n is a digit of pandigital.
A truly pandigital number has a bitmask of bitsAll = (1 << 9) | (1 << 8) | ... | (1 << 2) | (1 << 1).
According to the problem statement, a digit must not be zero. That means the lowest bit is always 0:
bitsAll_9=1022 (and bitsAll_8=510 for alternate Hackerrank input).

My program process all numbers i from 2 to maxFactor and multiplies them with 1, 2, ... until the concatenated result's bitsUsed == bitsAll.

Each product = i * multiplier (where multiplier = 1, 2, ...) is split into its digits and each digit added to the bitmask bitsUsed.
When a collision occurs (that digit was already used before), then bitsUsed is set to an invalid value and the next number i can be processed.

Modifications by HackerRank

The result may have either 8 or 9 digits and there is an upper limit for the start value.

Note

maxFactor cannot exceed 10000 because then i*1+i*2 will have more than 9 digits.
A few things are probably much easier when converting all numbers to std::string - at the cost of speed.

My code

… was written in C++ and can be compiled with G++, Clang++, Visual C++. You can download it, too.

The code contains #ifdefs to switch between the original problem and the Hackerrank version.
Enable #ifdef ORIGINAL to produce the result for the original problem (default setting for most problems).

#include <iostream>
 
int main()
{
unsigned int maxFactor, maxDigit; // 10000 and 9 for the original problem
std::cin >> maxFactor >> maxDigit;
 
// bitmask if all digits are used
unsigned int bitsAll = 0;
for (unsigned int i = 1; i <= maxDigit; i++)
bitsAll |= 1 << i;
 
#define ORIGINAL
#ifdef ORIGINAL
// largest pandigital number found so far
unsigned int largest = 0;
#endif
 
// try all numbers
for (unsigned int i = 2; i <= maxFactor; i++)
{
// the whole pandigital number
unsigned int pandigital = 0;
 
// multiply i by 1,2,3,...
unsigned int multiplier = 1;
 
// bitmask of used digits (nth bit is set if pandigital contains digit n)
unsigned int bitsUsed = 0;
 
while (bitsUsed < bitsAll)
{
// next step
unsigned int product = i * multiplier;
 
// extract right-most digit
while (product > 0)
{
// extract right-most digit
unsigned int digit = product % 10;
// remove it
product /= 10;
 
// make room to add i*multiplier lateron
pandigital *= 10;
 
// remember all digits we used
unsigned int bitMask = 1 << digit;
// we already had this digit ?
if (digit == 0 || (bitsUsed & bitMask) != 0)
{
bitsUsed = bitsAll + 1; // set to an invalid value
break;
}
 
// mark current digit as "used"
bitsUsed |= bitMask;
}
 
// keep going in the sequence
pandigital += i * multiplier;
multiplier++;
}
 
// enough digits generated ?
if (bitsUsed == bitsAll)
{
#ifdef ORIGINAL
if (largest < pandigital)
largest = pandigital;
#else
std::cout << i << std::endl;
#endif
}
}
 
#ifdef ORIGINAL
std::cout << largest << std::endl;
#endif
 
return 0;
}

This solution contains 15 empty lines, 16 comments and 9 preprocessor commands.

Interactive test

You can submit your own input to my program and it will be instantly processed at my server:

Input data (separated by spaces or newlines):

This is equivalent to
echo "200 9" | ./38

Output:

(please click 'Go !')

(this interactive test is still under development, computations will be aborted after one second)

Benchmark

The correct solution to the original Project Euler problem was found in less than 0.01 seconds on a Intel® Core™ i7-2600K CPU @ 3.40GHz.
(compiled for x86_64 / Linux, GCC flags: -O3 -march=native -fno-exceptions -fno-rtti -std=c++11 -DORIGINAL)

See here for a comparison of all solutions.

Note: interactive tests run on a weaker (=slower) computer. Some interactive tests are compiled without -DORIGINAL.

Changelog

February 25, 2017 submitted solution
April 12, 2017 added comments

Hackerrank

see https://www.hackerrank.com/contests/projecteuler/challenges/euler038

My code solves 6 out of 6 test cases (score: 100%)

Difficulty

Project Euler ranks this problem at 5% (out of 100%).

Hackerrank describes this problem as easy.

Note:
Hackerrank has strict execution time limits (typically 2 seconds for C++ code) and often a much wider input range than the original problem.
In my opinion, Hackerrank's modified problems are usually a lot harder to solve. As a rule thumb: brute-force is never an option.

Links

projecteuler.net/thread=38 - the best forum on the subject (note: you have to submit the correct solution first)

Code in various languages:

Python: www.mathblog.dk/project-euler-38-pandigital-multiplying-fixed-number/ (written by Kristian Edlund)
Haskell: github.com/nayuki/Project-Euler-solutions/blob/master/haskell/p038.hs (written by Nayuki)
Java: github.com/nayuki/Project-Euler-solutions/blob/master/java/p038.java (written by Nayuki)
Mathematica: github.com/nayuki/Project-Euler-solutions/blob/master/mathematica/p038.mathematica (written by Nayuki)
C: github.com/eagletmt/project-euler-c/blob/master/30-39/problem38.c (written by eagletmt)
Go: github.com/frrad/project-euler/blob/master/golang/Problem038.go (written by Frederick Robinson)
Javascript: github.com/dsernst/ProjectEuler/blob/master/38 Pandigital multiples.js (written by David Ernst)
Scala: github.com/samskivert/euler-scala/blob/master/Euler038.scala (written by Michael Bayne)

Heatmap

green problems solve the original Project Euler problem and have a perfect score of 100% at Hackerrank, too.
yellow problems score less than 100% at Hackerrank (but still solve the original problem).
gray problems are already solved but I haven't published my solution yet.
blue problems are solved and there wasn't a Hackerrank version of it at the time I solved it or I didn't care about it because it differed too much.

Please click on a problem's number to open my solution to that problem:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50
51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250
The 163 solved problems had an average difficulty of 22.2% at Project Euler and I scored 11,907 points (out of 13200) at Hackerrank's Project Euler+.
My username at Project Euler is stephanbrumme while it's stbrumme at Hackerrank.
more about me can be found on my homepage, especially in my coding blog.
some names mentioned on this site may be trademarks of their respective owners.
thanks to the KaTeX team for their great typesetting library !