Swift parse string with different formats

Question

I am working on a String parser in which the input can have various formats, and I don't know in advance which format is being used, so I need to write something that is flexible.

The first step is to check the first few characters, I can check that by using eg:

func parse(input: String) -> String {

   let result: String

   if (input.hasPrefix("foo") {
     result = doFoo(input)
   }
   else if (input.hasPrefix("bar") {
     result = doBar(input)
   }
   else if (input.hasPrefix("baz") {
     result = doBaz(input)
   }
   else {
     result = doBasic(input)
   }

   return result
}

and every doXXX() function has it's own parsing code, which again could have multiple options, such as different delimiters, etc.

This could easily turn into lots of if-else code, and I am wondering if with Swift there is a simpler way to do this. Maybe using switch-case statements, or something else? Could I use an enum for this?

EDIT: the code is inside a String extension.

marc_s · Accepted Answer · 2025-01-25 19:13:20Z

2

Here's how I would do it:

// This pattern matching operator defines what it means to have a
// closure as a pattern.  If the closure evaluates to true when called
// with `value` as an arg, then the `pattern` matches the `value`.
func ~=<T>(pattern: (T) -> Bool, value: T) -> Bool {
    return pattern(value)
}

// This type alias is just here to make the next line a bit more readable.
// A `BoolInstanceMethod<T, U>` is a closure type that represents an unapplied
// instance method that ultimately returns a Bool.

// For example, `String.hasPrefix` has type `(String) -> (String) -> Bool`.
// The first argument, of type `T` (String, in this case) is the instance
// this method will be called on.

// Say we call this: String.hasPrefix("The quick brown fox").
// The result has type `(String) -> Bool`.
// It's equivalent to "The quick brown fox".hasPrefix.

// We then call the resulting closure with the argument to hasPrefix
// For example: String.hasPrefix("The quick brown fox")("The")
// This has type `Bool`. It's the same as: "The quick brown fox".hasPrefix("The)
typealias BoolInstanceMethod<T, U> = (_ instance: T) -> (_ arg: U) -> Bool

// This function wraps a given instance method, in such a way as to reverse the
// order of the curried arguments. The given instance method is usually called as:
// Type.instanceMethod(instance)(arg), but this function allows you to swap it, to
// call it as: apply(Type.instanceMethod)(arg)(instance)
func apply<T, U>(instanceMethod: @escaping BoolInstanceMethod<T, U>) -> (_ arg: U) -> (_ instance: T) -> Bool {
    return { arg in
        return { instance in
            return instanceMethod(instance)(arg)
        }
    }
}

// Dummy functions to satisfy the compiler
func doFoo(_: String) -> String { return "" }
func doBar(_: String) -> String { return "" }
func doBaz(_: String) -> String { return "" }
func doBasic(_: String) -> String { return "" }

func parse(input: String) -> String {
    let result: String
    
    // The predicate of choice is made, in this case, String.hasPrefix.
    let hasPrefix = apply(instanceMethod: String.hasPrefix)
    
    // The switch calls `~=` for every case, giving it hasPrefix(...) and "input"
    // as args. The first case that makes `~=` yield `true` is executed.
    switch input {
    case hasPrefix("foo"): result = doFoo(input)
    case hasPrefix("bar"): result = doBar(input)
    case hasPrefix("baz"): result = doBaz(input)
    default: result = doBasic(input)
    }
    
    return result
}

// You could also implement parse like this:
func parse2(input: String) -> String {
    // You can save repeated application of the `input` parameter by doing it
    // just once at the end (see the `return` of this func).
    let action: (String) -> String
    
    // The predicate of choice is made, in this case, String.hasPrefix.
    let hasPrefix = apply(instanceMethod: String.hasPrefix)
    
    // The switch calls `~=` for every case, giving it hasPrefix(...) and "input"
    // as args. The first case that makes `~=` yield `true` is executed.
    
    switch input {
    case hasPrefix("foo"): action = doFoo
    case hasPrefix("bar"): action = doBar
    case hasPrefix("baz"): action = doBaz
    default: action = doBasic
    }
    
    return action(input)
}

edited Jan 25 at 19:13

marc_s

760k186 gold badges1.4k silver badges1.5k bronze badges

answered Dec 30, 2016 at 21:23

Alexander

63.9k13 gold badges106 silver badges177 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

koen Over a year ago

That looks promising. Unfortunately I get a compiler error for the first func: Member operator '~=' must have at least one argument of type 'String'; I also had to add the static keyword to that func. I have the parse code in a String extension, could that cause the error?

Alexander Over a year ago

I purposely made the ~= operator generic, so it applies to any types. In this case, it has to be a free standing function. If you really want to keep it in a String extension, then you could specialize it to take String instead of a generic T, but I advise against this.

koen Over a year ago

I changed that and it works now (I am using parse2).

koen Over a year ago

Yes, that's what I did :)

Collectives™ on Stack Overflow

Swift parse string with different formats

1 Answer 1

4 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

4 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related