March 11, 2017

* Featured Image – “Miss South Carolina Powerset Parse” by official_powerset, used under BY SA / Dropped Quality to 80% from original

I’ve run into situations where I had to parse a number into digits.

An easy way to do this is to convert a number into a string and returns a character array, and then convert each character into a number

When dealing with large sets of long numbers, it simply is not optimal.

There is a simple way to convert a number into digits, which requires a simple math.

Given an integer “val =213”, one might consider converting it to a string, breaking it to a character array and convert each character into an integer (it’s so simple in LINQ, it’s just tempting to implement it this way).

1 2 3 4 |
private static List<int> GetDigitsUsingConversion(int val) { return val.ToString().ToCharArray().Select(c => (int) c).ToList(); } |

The cost of type conversion from an integer to a string, to an array, and then back to integer is too high.

Now let’s take a look at another way using a simple math.

Given an integer 213, if you

- divide it by 1 and mod 10, you get 3
- divide it by 10 and mod 10, you get 1
- divide it by 100 and mod 10, you get 2

If you look at the returned result, it’s each digit returned in reverse order. There is a useful data structure, for holding data in reverse order, Stack.

An algorithm is fairly simple.

While the given number is greater than 0, divide it by 10^digit, put the digit into a stack, and lastly return the stack as a list.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
private static List<int> GetDigits(int val) { Stack<int> stack = new Stack<int>(); int number = val; while (number > 0) { var digit = number % 10; stack.Push(digit); number /= 10; } return stack.ToList(); } |

During each iteration, “number” is divided by 10 so it is equivalent to dividing by 10^digit.

I did a simple benchmarking (contrived but works for a simple demo) and the one requiring a type conversion to a string ran about 2x as long.

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 |
private const int UPTO = 1000000; public static void Main(string[] args) { int val = 123456789; Stopwatch watch = new Stopwatch(); watch.Start(); for (int i = 0; i < UPTO; i++) { List<int> digits = GetDigits(val); } watch.Stop(); Console.WriteLine("GetDigit took {0}ms", watch.ElapsedMilliseconds); watch.Start(); for (int j = 0; j < UPTO; j++) { List<int> digits2 = GetDigitsUsingConversion(val); } watch.Stop(); Console.WriteLine("GetDigitsUsingConversion took {0}ms", watch.ElapsedMilliseconds); } |

Result:

1 2 |
GetDigit took 754ms GetDigitsUsingConversion took 1468ms |

The source code is available on GitHub.

By using simple math, you can extract digits from a number.

It requires no type conversion thus saving run time.