Java 自定义比较器实现

Question

I am writing a Comparator with which I can sort a String array using character count. The Strings are all lowercase ASCII characters.

Eg.

input = {"art", "bash", "tar"};
Expected sorted output = {"art", "tar", "bash"}; 

Here, it does not matter if "bash" is at the beginning, but "art" and "tar" should be one after the other, in no particular order as they have matching character count.

I have this code, where I am using a simple array to check and keep the character count and compare.

Comparator<String> comparator = new Comparator<String>() {
    
    @Override
    public int compare(String s1, String s2){
        if (s1.equals(s2)) return 0;
        
        int[] a1 = getCountArray(s1);
        int[] a2 = getCountArray(s2);
        if (Arrays.equals(a1, a2)) return 0; // character count matches

        return s1.compareTo(s2);
    }

    private int[] getCountArray(String s) {
        int[] array = new int[256];
        for (int i=0; i<s.length(); i++){
            array[s.charAt(i) - 'a'] ++;
        }
        return array;
    }

};
String[] input = {"art", "bash", "tar"};
Arrays.sort(input, comparator); 

However, it does not work. What am I doing wrong?

Answer 1

Your comparator is not consistent:

• when comparing "art" against "bash" it returns that "art" is smaller than "bash"
• when comparing "bash" against "tar" it returns "bash" is smaller than "tar"
• therefore when comparing "art" against "tar" it must return "art" is smaller than "tar" (since "art" < "bash" < "tar"), but your comparator returns "art" is the same as "tar".

When you call Arrays.sort(input, comparator); the sort method will do pair-wise comparisons and will always find that "bash" has to be before "tar".

To fix this you must replace return s1.compareTo(s2) with something that compares the character count arrays a1 and a2 .