Java, writing my own split string method

Question

I need to be able to write my own split string method so that input like

String[] test1 = mySplit("ab#cd#efg#", "#");
System.out.println(Arrays.toString(test1));

will print [ab, #, cd, #, efg, #] to the console. So far I've got it to split like that but my way leaves awkward spaces where 2 delimiters are in a row, or a delimiter is at the start of the input.

public static String[] mySplit(String str, String regex)
{
    String[] storeSplit = new String[str.length()];
    char compare1, compare2;
    int counter = 0;

    //Initializes all the string[] values to "" so when the string
    //and char concatonates, 'null' doesn't appear.
    for(int i=0; i<str.length(); i++) {
        storeSplit[i] = "";
    }

    //Puts the str values into the split array and concatonates until
    //a delimiter is found, then it moves to the next array index.
    for(int i=0; i<str.length(); i++) {
        compare1 = str.charAt(i);
        compare2 = regex.charAt(0);

            if(!(compare1 == compare2)) {
                storeSplit[counter] += ""+str.charAt(i);
            } else {
                counter++;
                storeSplit[counter] = ""+str.charAt(i);
                counter++;
            }
    }
    return storeSplit;
}

When I use that method in my Test main, I get the output [ab, #, cd, #, efg, #, , , , ]. So I'm lost on how to fix the spacing of it all and I'll also need to be able to allow multiple delimiters which my code currently doesn't handle.

Also I know this code is really sloppy at the moment, just trying to lay down the concepts before the optimization.

Answer 1

The problem is straightforward, you have one offset walking through finding new matches (pos), and another showing then end of the last place you found a match (start).

public static String[] mySplit(String str, String regex)
{
    Vector<String> result = new Vector<String>;
    int start = 0;
    int pos = str.indexOf(regex);
    while (pos>=start) {
        if (pos>start) {
            result.add(str.substring(start,pos));
        }
        start = pos + regex.length();
        result.add(regex);
        pos = str.indexOf(regex,start); 
    }
    if (start<str.length()) {
        result.add(str.substring(start));
    }
    String[] array = result.toArray(new String[0]);
    return array;
}

This avoid extra looping and copies each character only once. Actually, because of the way that substring works, no characters are ever copied, only small string objects are created pointing to the original character buffer. No concatenation of strings is done at all, which is an important consideration.

Answer 2

I think your problem is that you are allocating storeSplit[] with a length that is longer than you need. If you are allowed to use ArrayList, use that to accumulate your results (and use the ArrayList.toArray() method to get the final return value for your function).

If you can't use ArrayList, then you will need to truncate your array before returning it (your counter variable will be of use in determining the correct length). To do that, you will need to allocate an array of correct length, then use System.arraycopy to populate it. Simpler to use ArrayList, but I don't know the exact requirements of your assignment.