Generally, is there a difference between
if("x".equalsIgnoreCase(myStrVar))
and
if(myStrVar.equalsIgnoreCase("x"))
in that the first will not cause a null pointer exception, but the second one will if in both scenarios myStrVar is null?
-
1Yes. Why? What happened to you when you tried it out?Kilian Foth– Kilian Foth2011-09-15 16:16:41 +00:00Commented Sep 15, 2011 at 16:16
-
1why don't you try it out in a junit test?hvgotcodes– hvgotcodes2011-09-15 16:16:43 +00:00Commented Sep 15, 2011 at 16:16
-
YOu've asked and answered the question already.Kal– Kal2011-09-15 16:16:56 +00:00Commented Sep 15, 2011 at 16:16
-
1I believe you're talking about "Yoda Conditions"mrkhrts– mrkhrts2011-09-15 16:20:30 +00:00Commented Sep 15, 2011 at 16:20
-
still +1 for the question, since the rule "you should start with the constant to avoid nullpointer dereferences" is quite useful.DaveFar– DaveFar2011-09-15 16:22:08 +00:00Commented Sep 15, 2011 at 16:22
Add a comment
|
5 Answers
Yes, that's correct. However, that doesn't mean you should always use the former ("safe") version.
If you are absolutely confident that myStrVal is non-null - meaning that for it to be null would indicate a bug - I would probably use the second form.
- I find it more readable (in the same way that I would write
if (x == 0)rather thanif (0 == x) - If the value is null, it's probably better to go bang as early as possible, rather than continuing while in an unexpected state
So:
- If you want to have special handling for
myStrValbeing null, then perform a check for that first - If you want to treat
myStrValbeing null in the same way as any other non-"x" value, use the first form - If you want to treat
myStrValbeing null as a bug, use the second form
Comments
If I got your question right, the answer is "yes". :)
1 Comment
Miserable Variable
His question is the answer. You have to reply in a question format :)
Your statement that the second version would cause a NullPointerException, whereas the first one wouldn't, is correct.
Comments
Obviously a NullPointerException will be raised in the second scenario if myStrVar == null. You can't call equalsIgnoreCase(String str) on null but you can pass in null as an argument.
Comments
Exactly, because the implementation is
public boolean equalsIgnoreCase(String anotherString) {
return (this == anotherString) ? true :
(anotherString != null) && (anotherString.count == count) &&
regionMatches(true, 0, anotherString, 0, count);
}
This is a nice instance of the (Yoda Conditions-)rule: "you should start with the constant to avoid nullpointer dereferences".
