Java HashSet中的重复项

Question

It seems like duplicates are allowed in HashSets. Why is this, how do I go about removing them, and why doesn't the second remove() work below? One method of removing all duplicates is new HashSet<>(set) , but is there a better way that doesn't involve creating a new object?

Set<ArrayList<String>> set = new HashSet<>();
ArrayList<String> a1 = new ArrayList<>();
ArrayList<String> a2 = new ArrayList<>();

a1.add("a");
set.add(a1);
a1.remove("a");

set.add(a2);

System.out.println(set.size());
System.out.println(set);

ArrayList<String> a3 = new ArrayList<>();
for (Object o : set) {
    boolean b = o.equals(a3) && (o.hashCode() == a3.hashCode());
    if (!b) System.out.println(false);
}

set.remove(new ArrayList<String>());
System.out.println(set);
set.remove(new ArrayList<String>());
System.out.println(set);
set.remove(set.iterator().next());
System.out.println(set);
System.out.println(set.iterator().next() == a1);

Output: set consists of two equal, empty lists, and the one that initially wasn't empty can't be removed.

2
[[], []]
[[]]
[[]]
[[]]
true

Answer 1

The location an element is stored in the HashMap depends on the hashCode of that element at the time it is added.

If after adding the element, you change a property of that element that causes its hashCode to change (in the case of an ArrayList element, removing an element from the list does exactly that), trying to find that element in the HashSet (or to remove it) will fail.

Answer 2

Hashing happens at insertion time for bucketing. If you change the object afterwards, its hashcode will change, but it will already be in its bucket. It will not be (directly) retrievable since you'll be trying to retrieve it with a hashcode different from the one you used to insert it.

a1.add("a"); 
set.add(a1); // hashed and bucketed
a1.remove("a"); // hash code changes but doesn't affect set

set.add(a2); // hashes to a different place than a1

Answer 3

If you modify the key of a Map or the element of a Set you are effectively corrupting it. There is no way for the collection to know you have changed the element, or to handle it correctly.

If you want to modify a key or element, you have to first remove it, modify it and add it back in.