doubleValue from NSString, decimals get removed

Question

Here I am using NSNumberFormater to set max number of decimals in a double to two. It seems to work fine, but when I try to convert it back to a double, from the string generated by NSNumberFormater, as you can see in the NSLog output, any decimal is removed and replaced by trailing zeros. Why is this?

- (double)randomDoubleBetween:(double)smallNumber and:(double)bigNumber {
    double diff = bigNumber - smallNumber;
    double ranNum = (((double) (arc4random() % ((unsigned)RAND_MAX + 1)) / RAND_MAX) * diff) + smallNumber;

    NSNumberFormatter *doubleValueWithMaxDecimalDigits = [[NSNumberFormatter alloc] init];
    [doubleValueWithMaxDecimalDigits setNumberStyle:NSNumberFormatterDecimalStyle];
    [doubleValueWithMaxDecimalDigits setMaximumFractionDigits:2];
    NSNumber *num = [NSNumber numberWithDouble:ranNum];

    NSString *valueToReturn = [doubleValueWithMaxDecimalDigits stringFromNumber:num];
    NSLog(@"String value %@", valueToReturn);

    NSLog(@"Double value converted from string %f", [valueToReturn doubleValue]);
    return [valueToReturn doubleValue];
}

NSLog output:

2013-08-09 02:25:05.031 GG[27947:907] String value 176,19

2013-08-09 02:25:05.035 GG[27947:907] Double value converted from string 176.000000

Answer 1

Looks like -[NSString doubleValue] isn't respecting the locale decimal separator; it expects a point, not a comma.

// US uses point as decimal separator
[doubleValueWithMaxDecimalDigits setLocale:[[NSLocale alloc]initWithLocaleIdentifier:@"en_US"]];

NSString *valueToReturn = [doubleValueWithMaxDecimalDigits stringFromNumber:num];

// Correctly includes fractional part
NSLog(@"Double value converted from string %lf", [valueToReturn doubleValue]);

// Italy uses a comma
[doubleValueWithMaxDecimalDigits setLocale:[[NSLocale alloc] initWithLocaleIdentifier:@"it_IT"]];
valueToReturn = [doubleValueWithMaxDecimalDigits stringFromNumber:num];

// Drops fractional part
NSLog(@"Double value converted from string %lf", [valueToReturn doubleValue]);

I'm not sure why you're formatting the number into a string that you don't use, but if you're going to continue to do that, use the same formatter again to parse the string into a double before returning.

(Aside, your format specifier is wrong; use %lf for double.)