Ultra fast Java string tokenizer

StringTokenizer.java

public class StringTokenizer {
    private static ThreadLocal<String[]> tempArray = new ThreadLocal<String[]>();

    public static String[] tokenize(String string, char delimiter) 
    {
        String[] temp = tempArray.get();
        int tempLength = (string.length() / 2) + 2;

        if (temp == null || temp.length < tempLength) 
        {
            temp = new String[tempLength];
            tempArray.set(temp);
        }


        int wordCount = 0;
        int i = 0;
        int j = string.indexOf(delimiter);

        while (j >= 0)
        {
            temp[wordCount++] = string.substring(i, j);
            i = j + 1;
            j = string.indexOf(delimiter, i);
        }

        temp[wordCount++] = string.substring(i);

        String[] result = new String[wordCount];
        System.arraycopy(temp, 0, result, 0, wordCount);
        return result;
    }
}

Fixed the ArrayIndexOutOfBounds

Why did you decide to use a ThreadLocal class variable for the tempArray instead of declaring it inside the function?

Reusing an object is faster. I can't recall if we actually proved it how ever //JT 2012/5/24 Yossale < [email protected]

Why did you decide to use a ThreadLocal class variable for the tempArray instead of declaring it inside the function? --- Reply to this email directly or view it on GitHub: https://gist.github.com/1056060

## ## Jonas Tärnström Product Manager • e-mail: [email protected] • skype: full name "Jonas Tärnström" • phone: +46 (0)734 231 552 ESN Social Software AB www.esn.me

Can you provide a version that supports multiple delim chars?

why/how is this ultra fast?