Number of occurrences of substring in string in Swift

My main string is "hello Swift Swift and Swift" and substring is Swift. I need to get the number of times the substring "Swift" occurs in the mentioned string.

This code can determine whether the pattern exists.

var string = "hello Swift Swift and Swift"

if string.rangeOfString("Swift") != nil {
    println("exists")
}

Now I need to know the number of occurrence.

A simple approach would be to split on "Swift", and subtract 1 from the number of parts:

let s = "hello Swift Swift and Swift"
let tok =  s.components(separatedBy:"Swift")
print(tok.count-1)

This code prints 3.

Edit: Before Swift 3 syntax the code looked like this:

let tok =  s.componentsSeparatedByString("Swift")

Should you want to count characters rather than substrings:

extension String {
    func count(of needle: Character) -> Int {
        return reduce(0) {
            $1 == needle ? $0 + 1 : $0
        }
    }
}

Optimising dwsolbergs solution to count faster. Also faster than componentsSeparatedByString.

extension String {
    /// stringToFind must be at least 1 character.
    func countInstances(of stringToFind: String) -> Int {
        assert(!stringToFind.isEmpty)
        var count = 0
        var searchRange: Range<String.Index>?
        while let foundRange = range(of: stringToFind, options: [], range: searchRange) {
            count += 1
            searchRange = Range(uncheckedBounds: (lower: foundRange.upperBound, upper: endIndex))
        }
        return count
    }
}

Usage:

// return 2
"aaaa".countInstances(of: "aa")

• If you want to ignore accents, you may replace options: [] with options: .diacriticInsensitive like dwsolbergs did.
• If you want to ignore case, you may replace options: [] with options: .caseInsensitive like ConfusionTowers suggested.
• If you want to ignore both accents and case, you may replace options: [] with options: [.caseInsensitive, .diacriticInsensitive] like ConfusionTowers suggested.
• If, on the other hand, you want the fastest comparison possible and you can guarantee some canonical form for composed character sequences, then you may consider option .literal and it will only perform exact matchs.