Double to Int Issue

[phosphide]

Hi all, I cannot believe I am having such an issue with this. I am trying to convert a double to an integer with the following code.

double yearsBetween = (date2.getTime() - date1.getTime())/365;

System.out.println("Years Between is: "+yearsBetween);

int roundUp = (int)yearsBetween;

System.out.println("Round up is = "+roundUp);

But the output looks like this:

Years Between is: 1.73195546301E11
Round up is = 2147483647

How the heck is that turning into 2147483647? Am I doing something wrong here? I know it will actually round down but I am going to add 1 which is why it is called round up. I also tried:

int roundUp = yearsBetween.intValue();

But I was getting a "Double cannot be dereferenced" error.

Any ideas?

[sunde887]

The "double cannot be dereferenced" stems from calling a method on a double, which don't exist for primitives. The other problem looks like you are putting a very large value into a smaller value, ints are 32 bits, so the max is about 2 billion, and -2 billion is the min.

Doubles are 64 bits and can hold much larger values. See if casting to long helps, it can hold 64 bits I believe, which (off the top of my head) I believe is ~9 trillion.

http://download.oracle.com/javase/tu...datatypes.html

[phosphide]

Uh, I'm not using any large numbers. The value of the double comes out to 1.73 and it should round to 1 when I convert it to an integer. It is somehow converting to 2147483647.

[Fubarable]

Hi all, I cannot believe I am having such an issue with this. I am trying to convert a double to an integer with the following code.

double yearsBetween = (date2.getTime() - date1.getTime())/365;

You know of course that Date#getTime() returns a result representing milliseconds, not days, and so dividing that difference by 365 will return a number that is not close to representing a year by a couple orders of magnitude.

And here,

int roundUp = yearsBetween.intValue();

as sunde states, you're trying to call a method on a double, a primitive. That's like trying to do this:

 int foo = 10.5.intValue();

which makes no sense whatsoever.

[sunde887]

Your output is: 1.73195546301E11 which is pretty large, unless I'm having a brain fart that number is equivalent to
173,195,546,301

[Fubarable]

Uh, I'm not using any large numbers. The value of the double comes out to 1.73 and it should round to 1 when I convert it to an integer. It is somehow converting to 2147483647.

LOL, look again. That's 1.73 x 10 to the 11th power! A very big number indeed.

[phosphide]

Oh for heavens sake, I should have realized that! I was totally reading that incorrectly. Thanks for the help yall!

[JosAH]

Oh for heavens sake, I should have realized that! I was totally reading that incorrectly. Thanks for the help yall!

According to the IEEE/754 (floating point formats) specification, when you cast a floating point number to an integer and it is too big, the resulting int value should be the largest representable int number; 2147483647 == 0x7FFFFFFF (this is Integer.MAX_VALUE).

kind regards,

Jos

[phosphide]

Well, the reason why I was screwing up my calculations was because I (for some reason) was looking at 1.73....e14 as 1.73 carried out to 14 decimals. When in fact that number was in scientific notation.

Though, your response does explain why I got the number that I did. Appreciate it.