Why can't custom comparator be used with hashSet to check duplicate objects

Issue

I want to use a custom comparator to filter duplicate objects from a HashSet.

In the Mobile class, I have defined a non-static nested Comparator class for comparing. The criterium which is used by this comparator is the id field of the Mobile object.

Set<Mobile> treeSet = new TreeSet<>(new Mobile().new Comparator()); // works


Set<Mobile> hashSet = new HashSet<>(new Mobile().new Comparator()); // does not work

Even though it works perfectly for a TreeSet, Eclipse shows a syntax error when I try this with a HashSet,

It appears that I need to override the equals and hashcode methods in the Mobile class if I want to use a HashSet. However, I would prefer to use a comparator instead.

Why is this the case?

Solution

HashSet is backed by a hashtable, which is an unordered data structure. This is why it has no need for a comparator. On the other hand, TreeSet is backed by a binary search tree (an ordered structure), and so it needs a comparator.

The correctness of a hashtable relies on the equals method, while its performance relies on a well-chosen hashCode. You can read more about it in the Javadoc of HashSet and the Object.hashCode() method itself.

Answered By - Marko Topolnik