Reading from text file with RegexMatch

First, a sample of the input data would be helpful for testing. That said:

• If you're using .NET 4 or greater you can use the File.ReadLines() method to read all the lines at once which may speed things up.
• Try to narrow the complexity of the regex pattern. Usages of * and + are going to be slower than other options. Or find a way to avoid regex all together.
• It looks like you're trying to match decimal numbers with the \d+.?\d* pattern, but the . character matches all characters. You may want the pattern \d+\.?\d*.
• Are you sure the input data is in the form of decimal colors? RGB implementations are almost always either bytes or integers.

One significant performance problem here is that you are doing two String operations on every line (well, almost every one). First, the line match, second the split. The second operation is also creating new String instances....

(included JoeClarks fix for the decimal point in the regex)

it is actually quite simple to bring this down to a single operation:

private static void GetOrigionalRGBColours(string txtFile)
{
    string tempLineValue;
    Regex regex = new Regex(@"^(\d+\.?\d*) (\d+\.?\d*) (\d+\.?\d*) SRGB$", RegexOptions.Compiled);

    using (StreamReader inputReader = new StreamReader(txtFile))
    {
        while (null != (tempLineValue = inputReader.ReadLine()))                     
        {
            Match match = regex.Match(tempLineValue);
            if (match.Success)
            {
                Decimal red = Convert.ToDecimal(match.Groups[1].Value);
                Decimal green = Convert.ToDecimal(match.Groups[2].Value);
                Decimal blue = Convert.ToDecimal(match.Groups[3].Value);
                if (    !(red == 1 && green == 1 && blue == 1)
                     && !(red == 0 && green == 0 && blue == 0))
                {
                    originalColourList.Add(new RGBColour(red, green, blue));
                }                    
            }
        }
    }
} 

I have run your code but it finishes for me for an input file of 290000 lines of 8MB under a second. Could you provide a sample of the file?

Update: My guess at this point is that the non matching lines which appear to be very long consume the most time. You could try to exclude them by checking for something simple like if the first character is a digit or if their length is below some threshold.

Can you provided some details on the implementation of RGBColour and what kind of collection originalColourList is. For my test I have assumed originalColourList is List<RGBColour> and RGBColour is a PODS like:

class RGBColour    {
    public decimal Red { get; set; }
    public decimal Green {get;  set;}
    public decimal Blue {get; set;}
}

I would imagine your while loop is slowing things down,

If you read in the complete file using File.ReadAllLines() then call Regex.Matches(), foreache'd through each match, it should be faster.

on a side note I personally like to store my large sql and regex filters/queries as a private const , obviously only if they ARE constant. though in this case it seems so.