Read UTF8 character in specify position from a NSS

Question

    NSString* str = @"1二3四5";
    NSLog(@"%c",[str characterAtIndex:0]); 
    NSLog(@"%c",[str characterAtIndex:1]);  

NSString - characterAtIndex works well on ASCII chars, but how could I get the UTF8 character at the index of 2?

-- updated --
It seems unichar(16bits) can't represent all the UTF8 encoding strings(8bites to 32bites), so are there any method to get the char from NSString?

Answer 1

Why don't you try to use something like that:

const char *yourWantedCharacter = [[yourSourceString substringWithRange:yourRange] UTF8String];

where yourSourceString is your NSString object, yourRange is an NSRange object with the index of the needed character as the location parameter and an length parameter of '0' (zero).

Answer 2

You'd use the more verbose methods:

NSRange rangeOfSecondCharacter = [str rangeOfComposedCharacterSequenceAtIndex:1];
NSString *secondCharacter = [str substringWithRange:rangeOfSecondCharacter];

...with proper bounds and range checking, of course. Note that this gives you an NSString, an object, not a unichar or some other primitive data type.

Answer 3

Unfortunately Dave's answer doesn't actually do what you want. The index supplied to rangeOfComposedCharacterSequenceAtIndex is an index of a UTF-16 code unit, 1 or 2 or which make a UTF-16 code point. So 1 is not the second UTF-16 code point if the first code point in the string requires 2 code units... (rangeOfComposedCharacterSequenceAtIndex returns the range of the code point which includes the code unit at the given index, so if your first char requires 2 code units then passing an index of 0 or 1 returns the same range).

If you want to find the UTF-8 sequence for a character you can use UTF8String and then parse the resultant bytes to find the byte sequence for the nth character. Or you can likewise use rangeOfComposedCharacterSequenceAtIndex starting at index 0 and iterate till you get to the nth character, then convert the 1 or 2 UTF-16 code units to UTF-8 code units.

I hope we're all missing something and this is built-in...

A start (needs bounds checking!) of a category which might help:

@interface NSString (UTF)

- (NSRange) rangeOfUTFCodePoint:(NSUInteger)number;

@end

@implementation NSString (UTF)

- (NSRange) rangeOfUTFCodePoint:(NSUInteger)number
{
    NSUInteger codeUnit = 0;
    NSRange result;
    for(NSUInteger ix = 0; ix <= number; ix++)
    {
        result = [self rangeOfComposedCharacterSequenceAtIndex:codeUnit];
        codeUnit += result.length;
    }
    return result;
}

@end

but this sort of stuff is more efficient using char * rather then NSString