3

I learned from looking at other references & SO that the Double has a lot of disadvantages.

Double disadvantages compared to double primitive data type:

  1. It is slower (reference : Java Double vs double: class type vs primitive type)
  2. It take more storage because of the metadata involved with it.

Now, I know there are some manipulation what we can do With Double. JavaDoc but most (85%) of these methods are static there fore it not a huge advantage. Another thing I know that double cannot be null.

Double doubleClass = null; // Is valid
double doubleType  = null; // Results in NPE

After all these disadvantages I cannot understand why we would use Double and not double in real life scenario. Can some give a real world example and explain.

Thanks

Community
  • 1
  • 1

3 Answers3

4

There are two main reasons as far as I can see.

  1. There are libraries which simply don't work with primitives. (The Collections API is the one you're most likely to encounter soon but anything that uses generics will be similar).
  2. Exactly the fact that Double is a reference type and therefore can be null. Sometimes that information is important. (Imagine a scenario when a user can fill in various fields in a form, some are optional and you want to find out whether a field was filled in. A null value is a reasonably good indicator that a form field was left empty.)

But in general you're right, whenever possible, you should use the primitive type.

biziclop
  • 48,926
  • 12
  • 77
  • 104
  • Thanks you this makes it much more clear. Appreciate the example that what I was looking for. –  Oct 10 '14 at 12:47
3

Double is a wrapper class over the primitive double. The Java Collection framework has containers like List, Set etc. The problem with those containers is that they cannot accept primitives. So, double is implicitly converted to Double and inserted into the collections (Autoboxing).

Double doubleClass = null; // Is valid --> Double is not a primitive. So, you can get NPE if you do doubleClass.someInstanceLevelFieldOrMethod.

double doubleType  = 5.5; //is a primitive

.

Also, Double will be slower than double because there is some book-keeping / overhead for Classes unlike primitives.

PS : Double provides many APi / methods to do things. example - Double.parseDouble(TypeX value) Creates a new double from value of TypeX

TheLostMind
  • 35,966
  • 12
  • 68
  • 104
  • Sorry if it sounds dumb but its main purpose for collections and not anywhere else ? –  Oct 10 '14 at 12:45
  • @newbie2314 - That is one of the reasons. Apart from that it provides methods [check here](http://docs.oracle.com/javase/7/docs/api/java/lang/Double.html). – TheLostMind Oct 10 '14 at 12:50
  • @TheLostMind Double.parseDouble isn't that again a static call ? – StackFlowed Oct 10 '14 at 14:38
  • @wrongAnswer - Yes. So it works if the instance is `null`. Thats why I mentioned - *you can get NPE if you do doubleClass.someInstanceLevelFieldOrMethod.* – TheLostMind Oct 10 '14 at 14:40
3

There are many main reasons that you should use Double

  1. The Collections API that need to use the class and not primitive type

  2. Some cases you know that double can have varying precision if the come from different data source. You can use the class .equals note most of the cases 

    double d1,d2; // Initalized to some value 
d1==d2; // will return true but some(very few will return false)

    You can read about this on .equals method

  3. As you have mentioned that Double can be null and double cannot be. Think of a scenario where a value of some object may or may exist if it doesn't then we want it as null in the database on retrieving it will throw an NPE with double but with Double it won't.

You should use the primitive type whenever possible and not Double.

StackFlowed
  • 6,664
  • 1
  • 29
  • 45