How to locate a sequence of values (specifically, bytes) within a larger collection in .NET

Question

I need to parse the bytes from a file so that I only take the data after a certain sequence of bytes has been identified. For example, if the sequence is simply 0xFF (one byte), then I can use LINQ on the collection:

byte[] allBytes = new byte[] {0x00, 0xFF, 0x01};
var importantBytes = allBytes.SkipWhile(byte b => b != 0xFF);
// importantBytes = {0xFF, 0x01}

But is there an elegant way to detect a multi-byte sequence - e.g. 0xFF, 0xFF - especially one that backtracks in case it starts to get a false positive match?

Answer 1

I'm not aware of any built-in way; as per usual, you can always write your own extension method. Here's one off the top of my head (there may be more efficient ways to implement it):

public static IEnumerable<T> AfterSequence<T>(this IEnumerable<T> source,
    T[] sequence)
{
    bool sequenceFound = false;
    Queue<T> currentSequence = new Queue<T>(sequence.Length);
    foreach (T item in source)
    {
        if (sequenceFound)
        {
            yield return item;
        }
        else
        {
            currentSequence.Enqueue(item);

            if (currentSequence.Count < sequence.Length)
                continue;

            if (currentSequence.Count > sequence.Length)
                currentSequence.Dequeue();

            if (currentSequence.SequenceEqual(sequence))
                sequenceFound = true;
        }
    }
}

I'll have to check to make sure that this is correct, but it should give you the basic idea; iterate through the elements, track the last sequence of values retrieved, set a flag when the sequence is found, and once the flag is set, start returning each subsequent element.

Edit - I did run a test, and it does work correctly. Here's some test code:

static void Main(string[] args)
{
    byte[] data = new byte[]
    {
        0x01, 0x02, 0x03, 0x04, 0x05,
        0xFF, 0xFE, 0xFD, 0xFC, 0xFB, 0xFA
    };
    byte[] sequence = new byte[] { 0x02, 0x03, 0x04, 0x05 };
    foreach (byte b in data.AfterSequence(sequence))
    {
        Console.WriteLine(b);
    }
    Console.ReadLine();
}

Answer 2

If you convert your bytes into a string, you can take advantage of the myriad of searching functions built into that, even if the bytes you're working with aren't actually characters in the traditional sense.

Answer 3

Just as a bit of theory; this is a regular language problem. You may be able to use a regular expression engine to detect it. The first google hit for "regular expression on stream" found

http://codeguru.earthweb.com/columns/experts/article.php/c14689