Converting Character and CodePoint in Swift

Question

Can I convert directly between a Swift Character and its Unicode numeric value? That is:

var i:Int = ...  // A plain integer index.
var myCodeUnit:UInt16 = myString.utf16[i]
// Would like to say myChar = myCodeUnit as Character, or equivalent.

or...

var j:String.Index = ... // NOT an integer!
var myChar:Character = myString[j]
// Would like to say myCodeUnit = myChar as UInt16

I can say:

myCodeUnit = String(myChar).utf16[0]

but this means creating a new String for each character. And I am doing this thousands of times (parsing text) so that is a lot of new Strings that are immediately being discarded.

Be aware that unicode is not a 16-bit charset, but a 21bit charset. — Deduplicator
– Deduplicator, Commented Jun 8, 2014 at 1:32
Can you do this operation in bulk? Say read a String of 1024 characters, then for in loop through each character instead of allocating a String per character? — Erik
– Erik, Commented Jun 8, 2014 at 3:04
Yes I am aware of the 21-bit issue. It is even more complicate than that, alas. Thanks SiLo, that is what I am doing. In fact, I keep an integer index and a String.Index going in parallel. But it seems a bit roundabout. — Andrew Duncan
– Andrew Duncan, Commented Jun 9, 2014 at 5:29

newacct · Accepted Answer · 2014-06-16 05:11:12Z

4

The type Character represents a "Unicode grapheme cluster", which can be multiple Unicode codepoints. If you want one Unicode codepoint, you should use the type UnicodeScalar instead.

answered Jun 16, 2014 at 5:11

newacct

123k29 gold badges168 silver badges227 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Daniel Chu · Accepted Answer · 2014-07-20 04:20:50Z

As per the swift book:

String to Code Unit

To get codeunit/ordinals for each character of the String, you can do the following:

var yourSwiftString = "甲乙丙丁"
for scalar in yourSwiftString.unicodeScalars {
    print("\(scalar.value) ")
}

Code Unit to String

Because swift current does not have a way to convert ordinals/code units back to UTF, the best way I found is to still NSString. i.e. if you have int ordinals (32bit but representing the 21bit codepoints) you can use the following to convert to Unicode:

var i = 22247
var unicode_str = NSString(bytes: &i, length: 4, encoding: NSUTF32LittleEndianStringEncoding)

Obviously if you want to convert a array of ints, you'll need to pack them into a array first.

Kyle · Accepted Answer · 2014-06-09 17:31:10Z

1

I spoke to an Apple engineer who is working on Unicode and he says they have not completed the implementation of unicode characters in strings. Are you looking at getting a code unit or a full character? Because the only and proper way to get at a full unicode character is by using a for each loop on a string. ie

for c in "hello" {
    // c is a unicode character of type Character
}

But, this is not implemented as of yet.

answered Jun 9, 2014 at 17:31

Kyle

214 bronze badges

1 Comment

Andrew Duncan Over a year ago

Thank you Kyle, it looks like the for-in iteration is working. I can't use that as-is because I need to break out of the loop and then back in, resuming where I left off, hundreds of times (tokenizing). BUT.. I see there is support for generators! I haven't figured out how to use them yet (perhaps not completely implemented) but that will be the way I go when I go. I do want to do this completely in Swift, just for consistency's sake. Otherwise I would get an NSData object.

Collectives™ on Stack Overflow

Converting Character and CodePoint in Swift

3 Answers 3

Comments

Comments

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

Comments

Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related