With the introduction of Swift I\'ve been trying to get my head round the new language
I\'m an iOS developer and would use types such as NSString, NSInteger, N
You should use the Swift native types whenever possible. The language is optimized to use them, and most of the functionality is bridged between the native types and the Foundation
types.
While String
and NSString
are mostly interchangeable, i.e, you can pass String
variables into methods that take NSString
parameters and vice versa, some methods seem to not be automatically bridged as of this moment. See this answer for a discussion on how to get the a String's length and this answer for a discussion on using containsString()
to check for substrings. (Disclaimer: I'm the author for both of these answers)
I haven't fully explored other data types, but I assume some version of what was stated above will also hold true for Array
/NSArray
, Dictionary
/NSDictionary
, and the various number types in Swift and NSNumber
Whenever you need to use one of the Foundation types, you can either use them to type variables/constants explicitly, as in var str: NSString = "An NSString"
or use bridgeToObjectiveC()
on an existing variable/constant of a Swift type, as in str.bridgeToObjectiveC().length
for example. You can also cast a String
to an NSString
by using str as NSString
.
However, the necessity for these techniques to explicitly use the Foundation types, or at least some of them, may be obsolete in the future, since from what is stated in the language reference, the String
/NSString
bridge, for example, should be completely seamless.
For a thorough discussion on the subject, refer to Using Swift with Cocoa and Objective-C: Working with Cocoa Data Types
Since the objective C types are still dynamically dispatched they're probably going to be slower. I'd say you're best served using the Swift native types unless you need to interact with objective-c APIs
Use the Swift native types whenever you can. In the case of String, however, you have "seamless" access to all the NSString
methods like this:
var greeting = "Hello!"
var len = (greeting as NSString).length
String gets revisions in swift 4. Now you can directly call count on it and it consider grapheme clusters as 1 piece, like an emoji. NSString is not updated and is counting it in another way.
var nsstring: NSString = "
Your best bet is to use Swift native types and classes, as some others have noted NSString has toll free translation to String, however, they're not the same a 100%, take for example the following
var nsstring: NSString = "\U0001F496"
var string: String = "\U0001F496"
nsstring.length
count(string)
you need to use the method count() to count the characters in string, also note that nsstring.length returns 2, because it counts its length based on UTF16.
Similar, YES The same, NO
Swift Strings are quite elegant and easy to use, unless you need to parse them. The whole concept of indexing into Swift strings is just plain crazy. Whenever I need to parse through a typical unicode string, I convert it to NSString. Swift makes this a bit tricky though in that the common "character" integer expressions like ' ' or 'A' or '0' don't work in Swift. You have to use 32, 65, 48. Unfortunately, I'm not kidding! Because of this, I've put most of my string parsing code into an NSString extension written in Objective-C.
Yes I do know WHY Swift's designers made String indexing so crazy: They wanted to be able to express many-byte characters like emoji as single "Characters". My choice would have been to let this rare use case be expressed as multiple UTF16 characters, but what the heck.