Question

What’s the difference between (fictious):

public Test GetTest()
{
  Object obj = new Test();
  return (Test)obj;
}

And

public Test GetTest()
{
  Object obj = new Test();
  return obj as Test;
}

Is it right that the first approach throws an exception if obj == null? And the second doesn’t?

Answer 1

No; both approaches will succeed and return null if obj is null. The difference lies in what happens if obj is not an instance of Test: the first approach will throw an exception, while the second one will succeed and return null.

In other words: Use the first approach if you know that your object is a Test or if you don’t know what it is, but you want an exception if it is not a Test. Use the second approach if you don’t know what your object is, but you just want a peaceful null if it is not a Test. You can also use as for type checking if you intend to do something special if the type check succeeds:

var t = obj as Test;
if (t != null) {
    ...
}

instead of:

if (obj is Test) {
    var t = (Test)obj;
    ...
}

In that way, you don’t need to repeat the type name, although the second form is probably clearer and avoids leaking t to the rest of the scope.

Also, see @il_guru’s post for some additional caveats related to as.