Meet swift regex

• Defined as struct Regex<Output>
• Can create with a literal /\d+/
  • Syntax is compatible with perl, python, ruby, java, nsregularexpression, etc
• Literal syntaxes built into compiler
• Can also create strings at runtime, which will throw errors at runtime if invalid
• Also can use swiftUI style regex Builder
• can use regex as a separator in string.split
  • also string.replace
• Interweave real parsers with regex
• Supports full unicode

Example

// CREDIT   02/01/2022  Payroll  $200.23
import RegexBuilder
let fieldSeparator = /\s{2,}|\t/
let transactionMatcher = Regex {
  Capture { /CREDIT|DEBIT/ }
  fieldSeparator
  One(.date(.numeric, locale: ..., timeZone: .gmt))
  fieldSeparator
  OneOrMore {
    NegativeLookahead { fieldSeparator }
    CharacterClass.any
  }
  fieldSeparator
  One(.localizedCurrency(code: "USD").locale(...))
}
// returns: Regex<(Substring, ...)>

• Extended delimeter. #/ .* / \d/#
  • Slashes without escapes
  • Whitespace is ignored, so can add linebreaks and alignment
• Capture labels (?<label>\d+) name the parameter in result tuple
• Can use unicode properties to identify character classes. Such as currencySymbol
• Match various unicode named characters, like /\N{SPARKLING_HEART}/
• Ignore case /blah/.ignoresCase()
• Can choose to match at unicode scaler level instead of character level.
• TryCapture can generalize captures with negative lookahead or any other custom capture or matching logic.
• Local{ /.../ } creates a scope where if the regex matches, any untried variations are discarded. Useful for paring down large failure search spaces. Disables global backtracking. Also known as an atomic noncapturing group.