5

An article on the Oracle Java Community sites gives as an example (for a JPA Converter, but that's not relevant, I guess) the following method:

public Boolean convertToEntityAttribute(String y) {
    String val = (String) y;
    if(val.equals("Y")){
        return true;
    } else {
        return false;
    }
}

What is the use of casting the String y to a String val? Is there a valid reason to do this?

Original article: What's New in JPA

7
  • 8
    return val.equals("Y") would suffice yes... Commented Dec 7, 2015 at 11:26
  • No need this redundant casting. Commented Dec 7, 2015 at 11:28
  • 1
    Mistake by a developer,like you and me. Thanks for Identifying :) Commented Dec 7, 2015 at 11:28
  • 2
    Tunaki - I'd of written return "Y".equals(val) Commented Dec 7, 2015 at 11:29
  • 2
    That whole sample code is awful. First of all, AttributeConverter needs two type parameters (as you'd expect, they are the two types you convert from/to), which the author neglects. As a result, the example doesn't even compile. Then there's the unnecessary cast and the icing on the cake is the if (boolean) {return true;} else {return false;} construct. Commented Dec 7, 2015 at 11:37

3 Answers 3

13

Such cast is completely unnecessary. I can imagine that it was before

public Boolean convertToEntityAttribute(Object y) {
    String val = (String) y;
    ...
}

But later the argument type was changed to String and the author simply forgot to remove the cast.

Sign up to request clarification or add additional context in comments.

Comments

4

Is there a valid reason to do this?

None whatsoever1.

But the flipside is that the Java compiler knows that the typecast is unnecessary and optimizes it away. So the only "damage" is to readability.

For example.

[stephen@blackbox tmp]$ cat Test.java 
public class Test {
    public void test (String x) {
        String s = (String) x;
        System.out.println(s);
    }
}
[stephen@blackbox tmp]$ javac Test.java 
[stephen@blackbox tmp]$ javap -c Test
Compiled from "Test.java"
public class Test {
  public Test();
    Code:
       0: aload_0
       1: invokespecial #1                  // Method java/lang/Object."<init>":()V
       4: return

  public void test(java.lang.String);
    Code:
       0: aload_1
       1: astore_2
       2: getstatic     #2                  // Field java/lang/System.out:Ljava/io/PrintStream;
       5: aload_2
       6: invokevirtual #3                  // Method java/io/PrintStream.println:(Ljava/lang/String;)V
       9: return
}
[stephen@blackbox tmp]$ 

The statement String s = (String) x; is compiled to a simple load and a store; no checkcast instruction.

I wouldn't be surprised if the JIT compiler was capable of optimizing away a redundant checkcast ... if it saw one.


1 - ... in hand written code. In source code that was generated, a redundant typecast could be serve the purpose of making it easier to write the source code generator. After all, the readability of generated code is largely irrelevant.

Comments

2

This redundant casting is useless.

Current code can be simplify to

public Boolean convertToEntityAttribute(String y) {
    return "Y".equals(y);
}

5 Comments

note that this is a different thing - code in the article would throw NPE on null values, whereas this would not. In some cases, NPE on null values is desired.
@eis both cases would not cause NPE, you can cast null to any reference type without getting any exception stackoverflow.com/questions/18723596/…
@Ian2thedv but you certainly can't call val.equals() on null value without getting an exception
@eis You will not get NPE by this. What are you referring ?
@RuchiraGayanRanaweera You will in the original code, so this is not equivalent to that.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.