Eliminate duplicates from strings

This can be simplified a great deal with an auxiliary data structure. Basically, we want to scan the string, and only add characters that we have not yet seen. Hence, if we use a data structure with fast insert/lookup, we can quickly tell if we've already seen that character without having to rescan over the string. In this case, this suggests a Set:

import java.util.Set;
import java.util.HashSet;

public static String removeDuplicate(String s)
{
    StringBuilder sb = new StringBuilder();
    Set<Character> seen = new HashSet<Character>();

    for(int i = 0; i < s.length(); ++i) {
        char c = s.charAt(i);
        if(!seen.contains(c)) {
            seen.add(c);
            sb.append(c);
        }
    }
    return sb.toString();
}

This trades time for space, so in this case, it should make the algorithm \$O(N)\$ time (as opposed to \$O(N^3)\$) at the cost of using \$O(N)\$ extra space.

Style

Java Code-style conventions put:

• the { brace at the end of the line that introduces the compound statement block, not on the next line.
• spaces between variables and operators: for (int i=0;i<length;i++)

Your code should look like:

public static String removeDuplicate(String s) {
    char[] temp = s.toCharArray();
    int length = temp.length;
    for (int i = 0; i < length; i++) {
        for (int j = i + 1; j < length; j++) {
            if (temp[i] == temp[j]) {
                int test = j;
                for (int k = j + 1; k < length; k++) {
                    temp[test] = temp[k];
                    test++;
                }
                length--;
                j--;
            }
        }
    }
    return String.copyValueOf(temp).substring(0, length);
}

Algorithm

You have three nested loops. This is inefficient.

A slight nit-pick is that you also use an inefficient way to get the resulting String using the copyValueOf(...) and then the substring(...).

I believe if you take an 'additive' approach to the process you will have a faster system. By extracting the onle loop in to a function it is also more readable:

private static final boolean foundIn(char[] temp, int size, char c) {
    for (int i = 0; i < size; i++) {
        if (temp[i] == c) {
            return true;
        }
    }
    return false;
}

public static String removeDuplicateY(String s) {
    char[] temp = s.toCharArray();
    int size = 0; //how many unique chars found so far
    for (int i = 0; i < temp.length; i++) {
        if (!foundIn(temp, size, temp[i])) {
            // move the first-time char to the end of the return value
            temp[size] = temp[i];
            size++;
        }
    }
    return new String(temp, 0, size);
}