NSNumbers and it's value representation issue

Question

Explanation Why this is not a duplicate of that pointed Q&A!:

I encountered this problem in relation to Objective-C minor implementation and NSNumber type has varying floating point precision and I just needed to know what really happens and what should I do to resolve this issue. That pointed Q&A doesn't suffice my question with its any answer.

Note: This is not regarding the types usage for the values. I was troubled to understand the behavior of NSNumbers.

I have this very simple method which uses NSNumber basically:

+ (NSNumber*)addNumber:(NSNumber*)firstNumber withSecondNumber:(NSNumber*)secondNumber {
    NSNumber *result = [NSNumber numberWithDouble:([firstNumber doubleValue] + [secondNumber doubleValue])];
    NSLog(@"Result: %@", result);   // This will print out our result for testing
    return result;
}

So if I call this method with parameters as follows:

[Fundamentals addNumber:[NSNumber numberWithDouble:5.0]
       withSecondNumber:[NSNumber numberWithDouble:3.5]];

Printed output: 8.5 ---> Exactly what I was expecting

But if I call this method with the following parameters:

[Fundamentals addNumber:[NSNumber numberWithDouble:50.0]
       withSecondNumber:[NSNumber numberWithDouble:36.54]];

It prints out the following: 86.53999999999999 ---> Which troubled me and still amazed me and questions my brain till it hurts! :(

Note: Not only when printing, even in debugger also it shows the value representation as it is printed out above.

Can somebody explain this please?

Thanks in advanced!

Have you tried to use [NSString stringWithFormat:@"%.1f", YourDoubleNumber] ? Should solve your precision — OhadM, Jul 05 '16 at 05:49
Nope! Actually I don't want to format the output as a string. I need the actual value representation to be a natural one! — Randika Vishman, Jul 05 '16 at 05:55
Thanks for down voting my question! But can somebody give an answer and why this should be down voted? — Randika Vishman, Jul 05 '16 at 06:27
This is a duplicate, no matter if it uses NSNumber instead of double. — Rudy Velthuis, Jul 05 '16 at 07:00
36.54 cannot be stored exactly in a `double`. `NSNumber` may be the interface of it, but internally, if you tell it to use `double` (`numberWithDouble:`), it uses `double` to store the value, and so what you want cannot be done. This is in fact the same old question once again. — Rudy Velthuis, Jul 05 '16 at 07:42

score 0 · Answer 1 · answered Jul 05 '16 at 07:02

To format float number, you can use NSNumberFormatter

NSNumberFormatter *formatter = [[NSNumberFormatter alloc] init];

[formatter setNumberStyle:NSNumberFormatterDecimalStyle];
[formatter setMaximumFractionDigits:2];
[formatter setRoundingMode: NSNumberFormatterRoundUp];

NSString *numberString = [formatter stringFromNumber:[NSNumber numberWithFloat:22.368511]];

NSLog(@"Result...%@",numberString);//Result 22.37

NSNumbers and it's value representation issue

1 Answers1