Double datatype, count decimals after decimal-place

Question

Method below should return a answer to "What decimal precision, quantity of trailing numbers, do this (double) value have?". I got it right when the values looks like 5900.43, 5900.043 and so on. When the method recieves 5900.00, it return 0, which is wrong (in my need).

   /// <summary>
   /// Return count of decimals after decimal-place 
   /// </summary>
   private int getDecimalCount(double val)
   {
        int count = 0;
        ...
        return count;
   }

When parameter 'val' is (samples of normal values to this method)

5900.00 return 2
5900.09 return 2
5900.000 return 3
5900.001 return 3
1.0 return 1
0.0000005 return 7
1.0000000 return 7
5900822 return 0

My problem is counting .0 .00 .000 and so on

Question:
How do I solve this? If this method isn't possible with doubles, how else?

[EDITED]
- Minor text change to make a more clear question,
- Language correction due to miss-used meaning of 'decimal places'

Answer 1

EDIT: Updated to simplify handling different cultures.

Similarly to @Miguel, here's how I would handle it:

public static int getDecimalCount(double dVal, string sVal, string culture)
{
    CultureInfo info = CultureInfo.GetCultureInfo(culture);

    //Get the double value of the string representation, keeping culture in mind
    double test = Convert.ToDouble(sVal, info);

    //Get the decimal separator the specified culture
    char[] sep = info.NumberFormat.NumberDecimalSeparator.ToCharArray();

    //Check to see if the culture-adjusted string is equal to the double
    if (dVal != test)
    {
        //The string conversion isn't correct, so throw an exception
        throw new System.ArgumentException("dVal doesn't equal sVal for the specified culture");
    }

    //Split the string on the separator 
    string[] segments = sVal.Split(sep);

    switch (segments.Length)
    {
        //Only one segment, so there was not fractional value - return 0
        case 1:
            return 0;
        //Two segments, so return the length of the second segment
        case 2:
            return segments[1].Length;

        //More than two segments means it's invalid, so throw an exception
        default:
            throw new Exception("Something bad happened!");
    }
}

And a shortcut method for US English:

public static int getDecimalCount(double dVal, string sVal)
{
    return getDecimalCount(dVal, sVal, "en-US");
}

Testing:

static void Main(string[] args)
{
    int i = 0;

    double d = 5900.00; 
    string s = "5900.00";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 2. Actual output: {0}", i);
    Console.WriteLine();

    d = 5900.09;
    s = "5900.09";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 2. Actual output: {0}", i);
    Console.WriteLine();

    d = 5900.000;
    s = "5900.000";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 3. Actual output: {0}", i);
    Console.WriteLine();

    d = 5900.001;
    s = "5900.001";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 3. Actual output: {0}", i);
    Console.WriteLine();

    d = 1.0; 
    s = "1.0";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 1. Actual output: {0}", i);
    Console.WriteLine();

    d = 0.0000005; 
    s = "0.0000005";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 7. Actual output: {0}", i);
    Console.WriteLine();

    d = 1.0000000; 
    s = "1.0000000";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 7. Actual output: {0}", i);
    Console.WriteLine();

    d = 5900822; 
    s = "5900822";
    Console.WriteLine("Testing with dVal = {0} and sVal = {1}.", d, s);
    i = getDecimalCount(d, s);
    Console.WriteLine("Expected output: 0. Actual output: {0}", i);

    Console.ReadLine();
}

And finally, the output of the test:

Testing with dVal = 5900 and sVal = 5900.00. Expected output: 2. Actual output: 2

Testing with dVal = 5900.09 and sVal = 5900.09. Expected output: 2. Actual output: 2

Testing with dVal = 5900 and sVal = 5900.000. Expected output: 3. Actual output: 3

Testing with dVal = 5900.001 and sVal = 5900.001. Expected output: 3. Actual output: 3

Testing with dVal = 1 and sVal = 1.0. Expected output: 1. Actual output: 1

Testing with dVal = 5E-07 and sVal = 0.0000005. Expected output: 7. Actual output: 7

Testing with dVal = 1 and sVal = 1.0000000. Expected output: 7. Actual output: 7

Testing with dVal = 5900822 and sVal = 5900822. Expected output: 0. Actual output: 0

Let me know if you have questions about it or it doesn't make sense.

Answer 2

This is not what has been asked, but it is what I needed. Maybe it will help someone:

 private static int getDecimalCount(double val)
 {
     int i=0;
     while (Math.Round(val, i) != val)
         i++;
     return i;
 }
double val1 = 5900.00; int count1 = getDecimalCount(val1);//result: 0
double val2 = 5900.09; int count2 = getDecimalCount(val2);//result: 2
double val3 = 5900.000; int count3 = getDecimalCount(val3);//result: 0
double val4 = 1.0; int count4 = getDecimalCount(val4);//result: 0
double val5 = 0.0000005; int count5 = getDecimalCount(val5);//result: 7
double val6 = 1.0000000; int count6 = getDecimalCount(val6);//result: 0
double val7 = 5900822; int count7 = getDecimalCount(val7);//result: 0

double val9 = 4.5565d; int count9 = getDecimalCount(val9);//result: 4

Answer 3

You're thinking of the String representation of a double. A double doesn't make a difference between these for example:

123
123.0
123.000000

Thats all the same double, how many decimal places you get, really depends on your String formatting options (or which are set by default). What you can do, is format your doubles as Strings with a lot of decimal places, then strip all trailing zeros, and then count the decimal places.

Answer 4

Create a method that receives a string instead of a double:

    public int CountDecimalDigits(string value)
    {
        char[] possibleChars = "0123456789.".ToCharArray();
        int decimalPoints = 0;
        foreach (char ch in value)
        {
            if (Array.IndexOf(possibleChars, ch) < 0)
                throw new Exception();
            if (ch == '.') decimalPoints++;
        }
        if (decimalPoints > 1)
            throw new Exception();
        if (decimalPoints == 0) return 0;
        return value.Length - value.IndexOf('.') - 1;
    }