Why can I use Strings as keys in a HashMap?

Question

If two Strings that are the same are not actually identical, then why can I use strings as keys in a HashMap without using the same String object?

String s1 = "Test";
String s2 = "Test";

System.out.println(s1 == s2); // should be false
System.out.println(s1.equals(s2)); // should be true

HashMap<String, String> map = new HashMap();
map.put(s1, "foo");
System.out.println(map.get(s2)); // should be "foo"--but why?

Does HashMap have some special behavior for String objects? If not, why can two "different" strings be used to put and to get values from a hash?

Answer 1

HashMap compares objects by calling equals() and hashCode().
String overrides these methods to compare by value.

Answer 2

In general, you can use String objects because HashMap uses equals() and not == to test for key equality.

Answer 3

If two Strings that are the same are not actually equal

But they are. They are equal under the equals() method, and that is the technique specified for equality testing in the Map interface.

System.out.println(s1 == s2); // should be false

But it isn't false! Both refer to the same string because of constant pooling by the compiler.

Answer 4

When the HashMap compares the key internally, it uses the equals() method, not ==. So object equality is fine for a key match, reference equality is not required if equals() is overridden (as in the case of java.lang.String.)