Extracting a Decimal from a string

Question

I have the following method which extracts a decimal value from a string.

using System;
using System.Linq;
using System.Text;

namespace Foo.Bar.Common.Converters
{
    public static class DecimalConverter
    {
        public static Decimal ExtractDecimalFromString(string str)
        {
            var sb = new StringBuilder();
            foreach (var c in str.Where(c => c == '.' || Char.IsDigit(c)))
            {
                sb.Append(c);
            }
            return Convert.ToDecimal(sb.ToString());
        }
    }
}

Is this the most effecticve way to do it?

Answer 1

Regular expressions can be a good way of processing text input in situations like this. Your current process finds every digit in the input string, concatenates it together, and then parses the result.

I doubt that this is a good solution because it makes examples like: "There are 10 people arriving at 9pm" parse as 109. Is that what you want?

Even then, Regular expressions are probably a good solution:

str = Regex.Replace(str, @"[.\D+]", "");

In str, replace all non-digits with "" (i.e. remove any non-digits).

Your code will still fail with an exception for input like "Hello", which has no digits, because the ToDecimal call will fail. Values with multiple . characters will also be interesting...

Additionally, your code does not support negative input values. This can be a trick to accomplish, but, again, using regular expressions, it is not horrendous.

Finally, I feel you should only be parsing the first set of decimal-like digis in the input, not all of them.

Putting this all together, I would have:

• use regex
• parse the first set of digits
• accept negative inputs

And use the code:

    public static Decimal ExtractDecimalFromString(string str)
    {

        Regex digits = new Regex(@"^\D*?((-?(\d+(\.\d+)?))|(-?\.\d+)).*");
        Match mx = digits.Match(str);
        //Console.WriteLine("Input {0} - Digits {1} {2}", str, mx.Success, mx.Groups);

        return mx.Success ? Convert.ToDecimal(mx.Groups[1].Value) : 0;
    }

I have put this, with some test cases, in this Ideone here;

Answer 2

Assuming you only want the first decimal embedded in the string and that your data is strongly typed and formatted, you should be able to short cut the loop by stopping after you extract the number. Something like this:

public static Decimal ExtractDecimalFromString(string str)
{
    var sb = new StringBuilder();
    foreach (var c in str.SkipWhile(c => c != '.' && !Char.IsDigit(c)))
    {
        if(c != '.' && !Char.IsDigit(c))
        {
            return Convert.ToDecimal(sb.ToString());
        }
        sb.Append(c);
    }
    return Convert.ToDecimal(sb.ToString());   
}