Arrays and Strings - Beginning Java 2 SDK - page 140

String Interning
Having convinced you of the necessity for using the equals method for comparing strings, we can now
reveal that there is a way to make comparing strings with the == operator effective. The mechanism to
make this possible is called string interning . String interning ensures that no two String objects
encapsulate the same string so all String objects encapsulates unique strings. This means that if two
String variables reference strings that are identical, the references must be identical too. To put it
another way, if two String variables contain references that are not equal, they must refer to strings
that are not equal. So how do we arrange that all String objects encapsulate unique strings? You just
call the intern() method for every new String object that you create. For instance, let's amend a bit
of an earlier example:
String string1 = "Too many ";
String string2 = "cooks";
String string3 = "Too many cooks";
// Make string1 and string3 refer to separate strings that are identical
string1 += string2;
string1 = string1.intern(); // Intern string1
The intern() method will check the string referenced by string1 against all the String objects
currently in existence. If it already exists, the current object will be discarded and string1 will contain
a reference to the existing object encapsulating the same string. As a result, the expression string1 ==
string3 will evaluate to true , whereas without the call to intern() it evaluated to false .
All string constants and constant String expressions are automatically interned. Thus if you add
another variable to the code fragment above:
String string4 = "Too " +"many ";
the reference stored in string4 will be automatically the same as the reference stored in string1 .
Only String expressions involving variables need to be interned. We could have written the statement
that created the combined string to be stored in string1 with the statement:
string1 = (string1 + string2).intern();
This now interns the result of the expression (string1 + string2) , ensuring that the reference
stored in string1 will be unique.
String interning has two benefits. First, it reduces the amount of memory required for storing String objects
in your program. If your program generates a lot of duplicate strings then this will be significant. Second, it
allows the use of == instead of the equals() method when you want to compare strings for equality. Since
the == operator just compares two references, it will be much faster than the equals() method, which
involves a sequence of character by character comparisons. This implies that you may make your program
run much faster, but only in certain cases. Keep in mind that the intern() method has to use the
equals() method to determine whether a string already exists. More than that, it will compare the current
string against a succession of, and possibly all, existing strings in order to determine whether the current
string is unique. Realistically you should stick to using the equals() method in the majority of situations
and only use interning when you are sure that the benefits outweigh the cost.