<< problem 40 - Champernowne's constant Coded triangle numbers - problem 42 >>

# Problem 41: Pandigital prime

We shall say that an n-digit number is pandigital if it makes use of all the digits 1 to n exactly once.
For example, 2143 is a 4-digit pandigital and is also prime.

What is the largest n-digit pandigital prime that exists?

# Algorithm

The largest pandigital number is 987654321. In order to find out whether a number x <= 987654321 is prime, I precompute all primes up to sqrt{987654321} approx 31426.
These "small" prime numbers will be kept in smallPrimes.

My second step is to generate all pandigital numbers: I create all permutations of the string "123456789" and perform a simple primality test (using smallPrimes).
Due to Hackerrank's variable number of digits, not only the 9-pandigitals numbers but also the 8-, 7-, 6-, ..., 2-pandigital numbers are checked, too.

The set panPrimes will contain all 2-, ..., 9-pandigital primes after those two precomputation steps.
Each test case look ups the closest bigger pandigital prime (upper_bound) and goes one step backwards.

## Note

When looking at the results I only saw 4- and 7-pandigital primes.
Modifying my loop in step 2 accordingly would provide a 10x speed-up.

# My code

… was written in C++11 and can be compiled with G++, Clang++, Visual C++. You can download it, too.

       #include <set>
#include <iostream>
#include <algorithm>

int main()
{
// precomputation step 1:
// find all primes below sqrt(987654321)
std::set<unsigned int> smallPrimes;
smallPrimes.insert(2);
for (unsigned int i = 3; i*i <= 987654321; i += 2)
{
bool isPrime = true;
for (auto p : smallPrimes)
{
// abort, no divisors possible
if (p*p > i)
break;

// divisor found ?
if (i % p == 0)
{
isPrime = false;
break;
}
}

// found a prime number
if (isPrime)
smallPrimes.insert(i);
}

// precomputation step 2:
// generate all permutations of the strings "12", "123", "1234", ..., "123456789"
// and test whether they are prime
std::set<unsigned int> panPrimes;
for (unsigned int digits = 2; digits <= 9; digits++)
{
std::string strNumber = "123456789";
// reduce number of digits
strNumber.erase(digits);

do
{
unsigned int number = std::stoi(strNumber);

// test whether pandigital number is prime
bool isPrime = true;
for (auto p : smallPrimes)
{
// abort, no divisors possible
if (p*p > number)
break;

// divisor found ?
if (number % p == 0)
{
isPrime = false;
break;
}
}

// found a pandigital prime ?
if (isPrime)
panPrimes.insert(number);
} while (std::next_permutation(strNumber.begin(), strNumber.end()));
}

// process input
unsigned int tests;
std::cin >> tests;
while (tests--)
{
unsigned int limit;
std::cin >> limit;

// find next larger pandigital prime number
auto i = panPrimes.upper_bound(limit);

// smaller than the smallest pandigital prime ?
if (i == panPrimes.begin())
{
std::cout << "-1" << std::endl;
continue;
}

// upper_bound() goes one step too far
i--;
// and print it
std::cout << *i << std::endl;
}

return 0;
}


This solution contains 13 empty lines, 18 comments and 3 preprocessor commands.

# Interactive test

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

Number of test cases (1-5):

Input data (separated by spaces or newlines):

This is equivalent to
echo "1 2500" | ./41

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 0.03 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

# Hackerrank

My code solved 5 out of 5 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.

projecteuler.net/thread=41 - 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-41-pandigital-prime/ (written by Kristian Edlund)
Java: github.com/nayuki/Project-Euler-solutions/blob/master/java/p041.java (written by Nayuki)
Mathematica: github.com/nayuki/Project-Euler-solutions/blob/master/mathematica/p041.mathematica (written by Nayuki)
C: github.com/eagletmt/project-euler-c/blob/master/40-49/problem41.cc (written by eagletmt)
Javascript: github.com/dsernst/ProjectEuler/blob/master/41 Pandigital prime.js (written by David Ernst)
Scala: github.com/samskivert/euler-scala/blob/master/Euler041.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 already 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
The 126 solved problems had an average difficulty of 16.0% at Project Euler and I scored 11,074 points (out of 12500) at Hackerrank's Project Euler+.
 << problem 40 - Champernowne's constant Coded triangle numbers - problem 42 >>
more about me can be found on my homepage.
some names mentioned on this site may be trademarks of their respective owners.
thanks to the KaTeX team for their great typesetting library !