Question

I think I don’t understand how the scope works in a switch case.

Can someone explain to me why the first code doesn’t compile but the second does ?

Code 1 :

 int key = 2;
 switch (key) {
 case 1:
      String str = "1";
      return str;
 case 2:
      String str = "2"; // duplicate declaration of "str" according to Eclipse.
      return str;
 }

Code 2 :

 int key = 2;
 if (key == 1) {
      String str = "1";
      return str;
 } else if (key == 2) {
      String str = "2";
      return str;
 }

How come the scope of the variable “str” is not contained within Case 1 ?

If I skip the declaration of case 1 the variable “str” is never declared…

Answer 1

I’ll repeat what others have said: the scope of the variables in each case clause corresponds to the whole switch statement. You can, however, create further nested scopes with braces as follows:

int key = 2;
switch (key) {
case 1: {
    String str = "1";
    return str;
  }
case 2: {
    String str = "2";
    return str;
  }
}

The resulting code will now compile successfully since the variable named str in each case clause is in its own scope.

Also note that the scope for case: (sans-{}) goes to the ‘global’ switch scope. And after that every other case scope (with or without {}) will receive those variables in a way uncommon for nested java scope, they can’t be overridden. So, for example, adding {} only to the second case would still give you the same error as it would get the variable from the first one injected, but adding braces to the first one would compile. I am not sure this is defined as such on the spec, but it what happens on most implementations.