That also implies (a,b) is not same as (b,a).

Test 5 only works if you disregard this and do the opposite.

    Set<String> set = new HashSet<>();
    for (int i = 0; i < t; i++) {
        String key = (pair_left[i].compareTo(pair_right[i]) <= 0) ? 
                     (pair_left[i] + " " + pair_right[i]) :
                     (pair_right[i] + " " + pair_left[i]);
        set.add(key);
        System.out.println(set.size());
    }

Test case 5th is incorrect: - Duplicate 191 / 259 ...

Test case 5 says there are 975 unique strings. I copied the data into Excel and ran this macro: =COUNTA(UNIQUE(D1:D1000)) The macro counted 990 unique strings.

I also ran the code in my IDE and it reported 990 unique strings, verifying Excel's findings.

It looks like this has been a known problem for a long time. I guess no one at HackerRank monitors this chat channel.

UPDATE: I copy/pasted the code from the Editorial tab. It also failed on test case 5.

It seems that the test #5 is not counting the pair on line 192 because it was added before but swapped. I guess that's an error on testing...