let startIndex = text.startIndex
var range = startIndex.advancedBy(1) ..< text.endIndex.advancedBy(-4)
let substring = text.substringWithRange(range)

你可以在这里看到完整的样本

注意:@airspeedswift对这种方法的权衡提出了一些非常深刻的观点，特别是隐藏的性能影响。字符串不是简单的野兽，获得特定的索引可能需要O(n)时间，这意味着使用下标的循环可能是O(n²)。我警告过你了。

你只需要添加一个新的下标函数，它接受一个范围，并使用advancedBy()走到你想要的地方:

import Foundation

extension String {
    subscript (r: Range<Int>) -> String {
        get {
            let startIndex = self.startIndex.advancedBy(r.startIndex)
            let endIndex = startIndex.advancedBy(r.endIndex - r.startIndex)

            return self[Range(start: startIndex, end: endIndex)]
        }
    }
}

var s = "Hello, playground"

println(s[0...5]) // ==> "Hello,"
println(s[0..<5]) // ==> "Hello"

(这绝对应该是语言的一部分。请dupe rdar://17158813)

为了好玩，你还可以在索引上添加一个+运算符:

func +<T: ForwardIndex>(var index: T, var count: Int) -> T {
  for (; count > 0; --count) {
    index = index.succ()
  }
  return index
}

s.substringWithRange(s.startIndex+2 .. s.startIndex+5)

(我还不知道这个是否应该成为语言的一部分。)

借鉴Rob Napier的经验，我开发了这些常见的字符串扩展，其中两个是:

subscript (r: Range<Int>) -> String
{
    get {
        let startIndex = advance(self.startIndex, r.startIndex)
        let endIndex = advance(self.startIndex, r.endIndex - 1)

        return self[Range(start: startIndex, end: endIndex)]
    }
}

func subString(startIndex: Int, length: Int) -> String
{
    var start = advance(self.startIndex, startIndex)
    var end = advance(self.startIndex, startIndex + length)
    return self.substringWithRange(Range<String.Index>(start: start, end: end))
}

用法:

"Awesome"[3...7] //"some"
"Awesome".subString(3, length: 4) //"some"

首先创建范围，然后是子字符串。你可以使用fromIndex..<toIndex语法如下:

let range = fullString.startIndex..<fullString.startIndex.advancedBy(15) // 15 first characters of the string
let substring = fullString.substringWithRange(range)

Rob Napier已经用下标给出了一个很棒的答案。但我觉得有一个缺点，因为没有检查出约束条件。这可能会导致崩溃。所以我修改了扩展，在这里

extension String {
    subscript (r: Range<Int>) -> String? { //Optional String as return value
        get {
            let stringCount = self.characters.count as Int
            //Check for out of boundary condition
            if (stringCount < r.endIndex) || (stringCount < r.startIndex){
                return nil
            }
            let startIndex = self.startIndex.advancedBy(r.startIndex)

            let endIndex = self.startIndex.advancedBy(r.endIndex - r.startIndex)

            return self[Range(start: startIndex, end: endIndex)]
        }
    }
}

下面的输出

var str2 = "Hello, World"

var str3 = str2[0...5]
//Hello,
var str4 = str2[0..<5]
//Hello
var str5 = str2[0..<15]
//nil

所以我建议总是检查if let

if let string = str[0...5]
{
    //Manipulate your string safely
}

由于String是NSString的桥接类型，“旧的”方法应该可以工作，但不清楚如何工作-例如，这也不起作用(似乎不是有效的语法): let x = str.substringWithRange(NSMakeRange(0,3))

对我来说，这是你问题中真正有趣的部分。String被桥接到NSString，所以大多数NSString方法直接作用于String。你可以自由地使用它们，不需要思考。例如，这正如你所期望的那样:

// delete all spaces from Swift String stateName
stateName = stateName.stringByReplacingOccurrencesOfString(" ", withString:"")

But, as so often happens, "I got my mojo workin' but it just don't work on you." You just happened to pick one of the rare cases where a parallel identically named Swift method exists, and in a case like that, the Swift method overshadows the Objective-C method. Thus, when you say str.substringWithRange, Swift thinks you mean the Swift method rather than the NSString method — and then you are hosed, because the Swift method expects a Range<String.Index>, and you don't know how to make one of those.

最简单的办法就是阻止霉霉这样做，明确地说:

let x = (str as NSString).substringWithRange(NSMakeRange(0, 3))

注意，这里没有涉及到重大的额外工作。“转换”并不意味着“转换”;String实际上是一个NSString。我们只是告诉Swift为了这一行代码的目的如何看待这个变量。

整个事情中真正奇怪的部分是导致所有这些麻烦的Swift方法是没有记录的。我不知道它的定义是什么;它不在NSString头文件中也不在Swift头文件中。