How to trim a String using Swift 3

Rather than mess around with trying to replace or remove the right characters or using regular expressions, I'd go with Foundation's built-in linguistic tagging support. It will do a lexical analysis of the string and return tokens of various types. Use it on this kind of string and it should reliably find any numbers in the string.

Something like:

var str = "from 3,95 €" let range = Range(uncheckedBounds: (lower: str.startIndex, upper: str.endIndex)) var tokenRanges = [Range<String.Index>]() let scheme = NSLinguisticTagSchemeLexicalClass let option = NSLinguisticTagger.Options() let tags = str.linguisticTags(in: range, scheme: scheme, options: option, orthography: nil, tokenRanges: &tokenRanges) let tokens = tokenRanges.map { str.substring(with:$0) } if let numberTagIndex = tags.index(where: { $0 == "Number" }) { let number = tokens[numberTagIndex] print("Found number: \(number)") } 

In this example the code prints "3,95". If you change str to "from £28.50", it prints "28.50".

One way is to place the unwanted strings into an array, and use String's replacingOccurrences(of:with:) method.

let stringToScan = "£28.50" let toBeRemoved = ["£", "€", "₹", "ab", "from"] var result = stringToScan toBeRemoved.forEach { result = result.replacingOccurrences(of: $0, with: "") } print(result) 

...yields "28.50".

If you just want to extract the numeric value use regular expression, it considers comma or dot decimal separators.

let string = "from 3,95 €" let pattern = "\\d+[.,]\\d+" do { let regex = try NSRegularExpression(pattern: pattern, options: []) if let match = regex.firstMatch(in: string, range: NSRange(location: 0, length: string.utf16.count)) { let range = match.range let start = string.index(string.startIndex, offsetBy: range.location) let end = string.index(start, offsetBy: range.length) print(string.substring(with: start..<end)) // 3,95 } else { print("Not found") } } catch { print("Regex Error:", error) } 

I asked if you had a fixed locale for this string, because then you can use the locale to determine what the decimal separator is: For example, try this in a storyboard.

let string = "some initial text 3,95 €" // define the string to scan // Add a convenience extension to Scanner so you don't have to deal with pointers directly. extension Scanner { func scanDouble() -> Double? { var value = Double(0) guard scanDouble(&value) else { return nil } return value } // Convenience method to advance the location of the scanner up to the first digit. Returning the scanner itself or nil, which allows for optional chaining func scanUpToNumber() -> Scanner? { var value: NSString? guard scanUpToCharacters(from: CharacterSet.decimalDigits, into: &value) else { return nil } return self } } let scanner = Scanner(string: string) scanner.locale = Locale(identifier: "fr_FR") let double = scanner.scanUpToNumber()?.scanDouble() // -> double = 3.95 (note the type is Double?) 

Scanners are a lot easier to use than NSRegularExpressions in these cases.

You can filter by special character by removing alphanumerics.

extension String { func removeCharacters(from forbiddenChars: CharacterSet) -> String { let passed = self.unicodeScalars.filter { !forbiddenChars.contains($0) } return String(String.UnicodeScalarView(passed)) } } let str = "£€₹jetztabfromnow12" let t1 = str.removeCharacters(from: CharacterSet.alphanumerics) print(t1) // will print: £€₹ let t2 = str.removeCharacters(from: CharacterSet.decimalDigits.inverted) print(t2) // will print: 12 

Updated 1:

var str = "£3,95SS" str = str.replacingOccurrences(of: ",", with: "") let digit = str.removeCharacters(from: CharacterSet.decimalDigits.inverted) print(digit) // will print: 395 let currency = str.removeCharacters(from: CharacterSet.alphanumerics) print(currency) // will print: £ let amount = currency + digit print(amount) // will print: £3,95 

Update 2:

let string = "£3,95SS" let pattern = "-?\\d+(,\\d+)*?\\.?\\d+?" do { let regex = try NSRegularExpression(pattern: pattern, options: []) if let match = regex.firstMatch(in: string, range: NSRange(location: 0, length: string.utf16.count)) { let range = match.range let start = string.index(string.startIndex, offsetBy: range.location) let end = string.index(start, offsetBy: range.length) let digit = string.substring(with: start..<end) print(digit) //3,95 let symbol = string.removeCharacters(from: CharacterSet.symbols.inverted) print(symbol) // £ print(symbol + digit) //£3,95 } else { print("Not found") } } catch { print("Regex Error:", error) }