If you’ve ever looked at trying to compare and evaluate two strings in Objective-C, you’ll find a whole host of methods to help you out. What you won’t find, however, is a method that will actually return the substring (or substrings) that belongs to one but not the other of the two strings you’re comparing.

I spent some time trying to figure out how to do this last week for one of my apps, and finally decided the best way was to break the strings down into arrays and then compare the arrays by removing items that belonged to one not the other.

I chose a fairly simple equivalence test based on length, and then set out to determine which was the longer of the two, before removing all the objects of the shorter one from the longer one. Here’s how I did it:


-(NSString *)getDifference: (NSString *)aString and:(NSString *)anotherString {

int i = aString.length;
int j = anotherString.length;
NSString *result, *longest, *shortest;

if (i == j) {
result = @””;
return result;
}

if (i > j) {
longest = aString;
shortest = anotherString;
} else {
longest = anotherString;
shortest = aString;
}

NSArray *fa = [longest componentsSeparatedByString: @” ” ];
NSArray *sa = [shortest componentsSeparatedByString: @” “];
NSMutableArray *remainder = [NSMutableArray arrayWithArray:fa];
[remainder removeObjectsInArray:sa];
result = [remainder componentsJoinedByString:@” “];
return result;

}

Here’s an example of how to call the method:


NSString *t1 = @"Mary had a little lamb";
NSString *t2 = @"Mary had a little lamb yesterday";

NSString *changes = [self getDifference:t1 and t2];
NSLog(@"%@", changes); //"yesterday"

This method is pretty basic as it stands. For one thing, the length equivalence test will certainly fail under some conditions of change; for another, you’d need to extend it to return items that are in the shorter string but not the longer (though that wouldn’t be that difficult). Nonetheless, the core of the method could be used as a basis for something far more robust. For example, this vastly more complex set of match patch and diff libraries appear to be essentially based on the same principle of breaking strings down into NSMutableArray’s and adding and removing objects, but with a vast army of conditions and other checks built in to improve reliability.

Still, the basic method I’ve posted above is the core idea I’m using in the new system profiler feature coming in the 1.5 release of FastTasks 2 (see the screenshot above), and at least at this stage of beta testing, seems to be doing the job it was designed for quite nicely.