I recently read some documentation and some blog entries about Key-Value Coding
and found it extremely useful. I want to utilize it for my case, but I haven't been successful so far.
In my case, I have an NSMutableArray
containing CGPoint
's transformed to NSValue
's. I will be adding many points to this data array and I want to get the minimum/maximum x and y values. For example:
NSMutableArray *data = [[NSMutableArray alloc] init];
[data addObject:[NSValue valueWithCGPoint:CGPointMake(20.0f, 10.0f)]];
[data addObject:[NSValue valueWithCGPoint:CGPointMake(5.0f, -15.0f)]];
[data addObject:[NSValue valueWithCGPoint:CGPointMake(-5.0f, 20.0f)]];
[data addObject:[NSValue valueWithCGPoint:CGPointMake(15.0f, 30.0f)]];
// min X: -5.0, max X: 20.0
// min Y: -15.0, max Y: 30.0
I have two approaches so far. The first one is to store these four values in class instance variables and when a new object is added, compare it with those variables and update them if necessary. The second one involves a for-loop to to find the extremum values, however this approach would be inefficient.
If possible, I would like to use KVC to do this task and I believe that it would be a more general solution than those I have. In the future, I might also need to remove some of the objects from the array and that would make my first approach inapplicable and all I am left with would be the for-loop.
I tried to use some key paths, i.e. @"@max.x"
, @"@max.position.x"
but all I got is an NSUnknownKeyException
.
You could make it work if you use a category to tell NSValue
how to access the x
and y
value of a CGPoint
:
@interface NSValue (MyKeyCategory)
- (id)valueForKey:(NSString *)key;
@end
@implementation NSValue (MyKeyCategory)
- (id)valueForKey:(NSString *)key
{
if (strcmp([self objCType], @encode(CGPoint)) == 0) {
CGPoint p = [self CGPointValue];
if ([key isEqualToString:@"x"]) {
return @(p.x);
} else if ([key isEqualToString:@"y"]) {
return @(p.y);
}
}
return [super valueForKey:key];
}
@end
Then
CGFloat maxx = [[yourArray valueForKeyPath:@"@max.x"] floatValue];
actually works. But note that a simple loop will be much faster, in particular if you have to compute all 4 values @max.x
, @max.y
, @min.x
, @min.y
.