IFormatProvider scientific conversion from double to string - number of digits

Question

I have a problem converting from double to string.

I want to convert:

double value: 0.0772486324655191

String value: 0.0772486324655191

and if the length after the decimal point is more than 16 digits, I want it to be like this:

double value: 0.00063500244832493823

string value: 6.3500244832493823e-004

I tried converting it with the IFormatProvider pattern:

0.0000000000000000e000

But the result in the first case is

7.7248632465519100e-002

How can I get the number of digits in my double vector? Or better: how should I use the format provider correctly?

String specifier; CultureInfo culture; specifier = "0.0000000000000000e000"; culture = CultureInfo.CreateSpecificCulture("en-US"); Console.WriteLine(DoubleMirrored[0].ToString(specifier, CultureInfo.InvariantCulture)); 

Answer 1

To do this, you definitely need to create your own formatter.

To create a custom formatter, here is what you should know:
string.Format has the following overload: string.Format(IFormatProvider, string, object[]) , so you must create an IFormatProvider that will "provide" ICustomFormatter that will handle your custom formatting. The same class can be easily used for both interfaces.

Here is some code that does exactly what you describe:

 public class DoubleFormatter : IFormatProvider, ICustomFormatter { // Implementation of IFormatProvider: public object GetFormat(Type t) { if (t == typeof(ICustomFormatter)) { return this; } return null; } // Implementation of ICustomFormatter: public string Format(string format, object arg, IFormatProvider provider) { // Search for the custom "EE" format specifier: if (format == null || !format.StartsWith("EE")) return null; format = format.Substring(2); // Trim "EE" // Determine how many digits before we cutoff: int digits; if (!int.TryParse(format, out digits)) { throw new FormatException("Format must contain digits"); } // Get the value: (note, this will work for any numeric type) var value = Convert.ToDouble(arg); // Convert to string without using Exponential format: var output = value.ToString("0."+(new string('#',digits)), provider); // Determine how many digits are showing: (this part isn't culture-compatible) var length = output.Length - output.IndexOf("."); if (length <= digits) { return output; } else { return value.ToString("E"+format, provider); } } } 

And here is an example of using this code:

 var tests = new[]{ 0.0000055555, 0.00000555555555555555555, }; var formatter = new DoubleFormatter(); foreach (var t in tests){ var result = string.Format(formatter, "{0:EE15}", t); Console.WriteLine(result); }