What is the best way to find the digit at n position in a decimal number?

Question

Background

I'm working on a symmetric rounding class and I find that I'm stuck with regards to how to best find the number at position x that I will be rounding. I'm sure there is an efficient mathematical way to find the single digit and return it without having to resort to string parsing.

Problem

Suppose, I have the following (C#) psuedo-code:

var position = 3;
var value = 102.43587m;
// I want this no ↑ (that is 5)

protected static int FindNDigit(decimal value, int position)
{
    // This snippet is what I am searching for
}

Also, it is worth noting that if my value is a whole number, I will need to return a zero for the result of FindNDigit.

Does anyone have any hints on how I should approach this problem? Is this something that is blaringly obvious that I'm missing?

Answer 1

(int)(value * Math.Pow(10, position)) % 10

Answer 2

How about:

(int)(value * Math.Pow(10, position)) % 10

Basically you multiply by 10 ^ pos in order to move that digit to the one's place, and then you use the modulus operator % to divide out the rest of the number.

Answer 3

How about this:

protected static int FindNDigit(decimal value, int position)
{
    var index = value.ToString().IndexOf(".");
    position = position + index;
    return (int)Char.GetNumericValue(value.ToString(), position);
}

Answer 4

Edited: Totally had the wrong and opposite answer here. I was calculating the position to the left of the decimal instead of the right. See the upvoted answers for the correct code.

Answer 5

using System;

public static class DecimalExtensions
{
    public static int DigitAtPosition(this decimal number, int position)
    {
        if (position <= 0)
        {
            throw new ArgumentException("Position must be positive.");
        }

        if (number < 0)
        {
            number = Math.Abs(number);
        }

        number -= Math.Floor(number);

        if (number == 0)
        {
            return 0;
        }

        if (position == 1)
        {
            return (int)(number * 10);
        }

        return (number * 10).DigitAtPosition(position - 1);
    }
}

Edit: If you wish, you may separate the recursive call from the initial call, to remove the initial conditional checks during recursion:

using System;

public static class DecimalExtensions
{
    public static int DigitAtPosition(this decimal number, int position)
    {
        if (position <= 0)
        {
            throw new ArgumentException("Position must be positive.");
        }

        if (number < 0)
        {
            number = Math.Abs(number);
        }

        return number.digitAtPosition(position);
    }

    static int digitAtPosition(this decimal sanitizedNumber, int validPosition)
    {
        sanitizedNumber -= Math.Floor(sanitizedNumber);

        if (sanitizedNumber == 0)
        {
            return 0;
        }

        if (validPosition == 1)
        {
            return (int)(sanitizedNumber * 10);
        }

        return (sanitizedNumber * 10).digitAtPosition(validPosition - 1);
    }

---

Here's a few tests:

using System;
using Xunit;

public class DecimalExtensionsTests
{
                         // digit positions
                         // 1234567890123456789012345678
    const decimal number = .3216879846541681986310378765m;

    [Fact]
    public void Throws_ArgumentException_if_position_is_zero()
    {
        Assert.Throws<ArgumentException>(() => number.DigitAtPosition(0));
    }

    [Fact]
    public void Throws_ArgumentException_if_position_is_negative()
    {
        Assert.Throws<ArgumentException>(() => number.DigitAtPosition(-5));
    }

    [Fact]
    public void Works_for_1st_digit()
    {
        Assert.Equal(3, number.DigitAtPosition(1));
    }

    [Fact]
    public void Works_for_28th_digit()
    {
        Assert.Equal(5, number.DigitAtPosition(28));
    }

    [Fact]
    public void Works_for_negative_decimals()
    {
        const decimal negativeNumber = -number;
        Assert.Equal(5, negativeNumber.DigitAtPosition(28));
    }

    [Fact]
    public void Returns_zero_for_whole_numbers()
    {
        const decimal wholeNumber = decimal.MaxValue;
        Assert.Equal(0, wholeNumber.DigitAtPosition(1));
    }

    [Fact]
    public void Returns_zero_if_position_is_greater_than_the_number_of_decimal_digits()
    {
        Assert.Equal(0, number.DigitAtPosition(29));
    }

    [Fact]
    public void Does_not_throw_if_number_is_max_decimal_value()
    {
        Assert.DoesNotThrow(() => decimal.MaxValue.DigitAtPosition(1));
    }

    [Fact]
    public void Does_not_throw_if_number_is_min_decimal_value()
    {
        Assert.DoesNotThrow(() => decimal.MinValue.DigitAtPosition(1));
    }

    [Fact]
    public void Does_not_throw_if_position_is_max_integer_value()
    {
        Assert.DoesNotThrow(() => number.DigitAtPosition(int.MaxValue));
    }
}