Return same double only if the double is an int? (no decimals) Obj-C

Question

I'm using a for-loop to determine whether the long double is an int. I have it set up that the for loop loops another long double that is between 2 and final^1/2. Final is a loop I have set up that is basically 2 to the power of 2-10 minus 1. I am then checking if final is an integer. My question is how can I get only the final values that are integers? My explanation may have been a bit confusing so here is my entire loop code. BTW I am using long doubles because I plan on increasing these numbers very largely.

for (long double ld = 1; ld<10; ld++) {
    long double final = powl(2, ld) - 1;
    //Would return e.g. 1, 3, 7, 15, 31, 63...etc.

    for (long double pD = 2; pD <= powl(final, 0.5); pD++) {
    //Create new long double
    long double newFinal = final / pD;
    //Check if new long double is int
    long int intPart = (long int)newFinal;
    long double newLong = newFinal - intPart;

    if (newLong == 0) {
        NSLog(@"Integer");
        //Return only the final ints?
    }
}
}

@nhgrif sorta has something. You should be able to take the number, cast to long, cast back to double, and then compare the two bit-for-bit. The only other thing you'd need to check is that the number of bits "used" in the long is fewer than the size of double's mantissa. — Hot Licks, Dec 12 '13 at 02:18
Oops -- I see you're using a `long double`, not casting between `long` and `double`. Perhaps you could cast to a `long long`? — Hot Licks, Dec 12 '13 at 02:20

nhgrif · Accepted Answer · 2013-12-12T02:31:25.297

Just cast it to an int and subtract it from itself?

long double d;
//assign a value to d

int i = (int)d;
if((double)(d - i) == 0) {
    //d has no fractional part
}

As a note... because of the way floating point math works in programming, this == check isn't necessarily the best thing to do. Better would be to decide on a certain level of tolerance, and check whether d was within that tolerance.

For example:

if(fabs((double)(d - i)) < 0.000001) {
    //d's fractional part is close enough to 0 for your purposes
}

You can also use long long int and long double to accomplish the same thing. Just be sure you're using the right absolute value function for whatever type you're using:

fabsf(float)
fabs(double)
fabsl(long double)

EDIT... Based on clarification of the actual problem... it seems you're just trying to figure out how to return a collection from a method.

-(NSMutableArray*)yourMethodName {
    NSMutableArray *retnArr = [NSMutableArray array];

    for(/*some loop logic*/) {
        // logic to determine if the number is an int

        if(/*number is an int*/) {
            [retnArr addObject:[NSNumber numberWithInt:/*current number*/]];
        }
    }

    return retnArr;
}

Stick your logic into this method. Once you've found a number you want to return, stick it into the array using the [retnArr addObject:[NSNumber numberWithInt:]]; method I put up there.

Once you've returned the array, access the numbers like this:

[[arrReturnedFromMethod objectAtIndex:someIndex] intValue];

Optionally, you might want to throw them into the NSNumber object as different types.

You can also use:

[NSNumber numberWithDouble:]
[NSNumber numberWithLongLong:]

And there are matching getters (doubleValue,longLongValue) to extract the number. There are lots of other methods for NSNumber, but these seem the most likely you'd want to be using.

Sorry but I already have this. I may have been unclear in my explanation. I know how to figure if its an int or not. Because d changes, I want to return only the d's that are ints. Thank you for the advice on checking if its 0. — Milo, Dec 12 '13 at 02:14
You're going to have to massively clarify where you're stuck. The `if` will return true for any double that doesn't have a fractional part and `false` for a double that does. Use that logic to do whatever you want now that it's sorting numbers out. — nhgrif, Dec 12 '13 at 02:19
Ok I have A. A is a double that has been ran through a loop and is now the numbers 1, 3, 7, 15, 31...etc. I now want to take A within its loop and divide it by another double, B, that has been ran through a loop. That double is 2 through A to the power of 1/2. We now have A/B. I have set A/B to the double C. Because all these values are changing I just want to NSLOG C when C is an int. Clear? — Milo, Dec 12 '13 at 02:21
Is this in a method? You need to return a collection of numbers back to the caller? — nhgrif, Dec 12 '13 at 02:22

Return same double only if the double is an int? (no decimals) Obj-C

1 Answers1