Swift Essentials

Numeric types in Swift can represent both signed and unsigned integral values with sizes 8, 16, 32, or 64 bits, as well as signed 32 or 64 bit floating point values. Numbers can include underscores to provide better readability; so 68_040 is the same as 68040:

> 3.141
$R0: Double = 3.141
> 299_792_458
$R1: Int = 299792458
> -1
$R2: Int = -1
> 1_800_123456
$R3: Int = 1800123456

Numbers can also be written in binary, octal, or hexadecimal using prefixes 0b, 0o (zero and the letter "o") or 0x. Note that Swift does not inherit C's use of a leading zero (0) to represent an octal value, unlike Java and JavaScript which do. Examples include:

> 0b1010011
$R0: Int = 83
> 0o123
$R1: Int = 83
> 0123
$R2: Int = 123
> 0x7b
$R3: Int = 123

There are three types of floating point values available in Swift, which use the IEEE754 floating point standard. The Double type represents 64 bits worth of data whilst Float stores 32 bits of data. In addition, Float80 is a specialized type that stores 80 bits worth of data.

By default, floating point values in Swift have the Double type. As floating point representation cannot represent some numbers exactly, some values will be displayed with a rounding error; for example:

> 3.141
$R0: Double = 3.141
> Float(3.141)
$R1: Float = 3.1400003

Floating point values can be specified in decimal or hexadecimal. Decimal floating point uses e as the exponent for base 10, whereas hexadecimal floating point uses p as the exponent for base 2. A value of AeB has the value A*10^B and a value of 0xApB has the value A*2^B:

> 299.792458e6
$R0: Double = 299792458
> 299.792_458_e6
$R1: Double = 299792458
> 0x1p8
$R2: Double = 256
> 0x1p10
$R3: Double = 1024
> 0x4p10
$R4: Double = 4096
> 1e-1
$R5: Double = 0.10000000000000001
> 1e-2
$R6: Double = 0.01
> 0x1p-1
$R7: Double = 0.5
> 0x1p-2
$R8: Double = 0.25
> 0xAp-1
$R9: Double = 5

Strings can contain escaped characters, unicode characters, and interpolated expressions. Escaped characters use a slash (\) and can be one of:

An interpolated string has an embedded expression which is evaluated, converted into a string, and concatenated into the result. These interpolated strings can capture local variables or expressions:

> "3+4 is \(3+4)"
$R0: String = "3+4 is 7"
> 3+4
$R1: Int = 7
> "7 x 2 is \($R1 * 2)"
$R2: String = "7 x 2 is 14"

Swift distinguishes between variables (which can be modified) and constants (which cannot be changed after assignment). Identifiers start with an underscore or alphabetic letter, followed by an underscore or alphanumeric character. In addition, other Unicode character points (such as emoji) can be used, although box lines and arrows are not allowed. Consult the Swift language guide for the full set of allowable Unicode characters. Generally, private use areas are not allowed and identifiers cannot start with a combining character (such as an accent).

Variables are defined with the var keyword and constants are defined with the let keyword. If the types are not specified, they are automatically inferred:

> let pi = 3.141
pi: Double = 3.141
> pi = 3
error: cannot assign to 'let' value 'pi'
> var i = 0
i: Int = 0
> ++i
$R0: Int = 1

Types can be explicitly specified. For example, to store a 32 bit floating point value, the variable can be defined as a Float, or to store a value as an unsigned 8 bit integer, UInt8:

> let e:Float = 2.718
e: Float = 2.71799994
> let ff:UInt8 = 255
ff: UInt8 = 255

To convert a number to a different type, it can be converted using the type initializer or assigned to a variable of a different type, provided it does not underflow or overflow:

> let ooff = UInt16(ff)
ooff: UInt16 = 255
> Int8(255)
error: integer overflows when converted from 'Int' to 'Int8'
Int8(255)
^
> UInt8(Int8(-1))
error: negative integer cannot be converted to unsigned type 'UInt8'
UInt8(Int8(-1))
^

Swift has two collection types: Array and Dictionary. They are strongly typed and generic, which ensures that the values of types assigned are compatible with the element type. The literal syntax for arrays uses [] to store a comma-separated list, while dictionaries use a comma-separated [key:value] format for entries. Collections defined with var are mutable; collections defined with let are immutable.

> var shopping = [ "Milk", "Eggs", "Coffee", ]
shopping: [String] = 3 values {
  [0] = "Milk"
  [1] = "Eggs"
  [2] = "Coffee"
}
> var costs = [ "Milk":1, "Eggs":2, "Coffee":3, ]
costs: [String : Int] = {
  [0] = { key = "Coffee" value = 3 }
  [1] = { key = "Milk"   value = 1 }
  [2] = { key = "Eggs"   value = 2 }
}

Arrays and dictionaries can be indexed using subscript operators, reassigned, and added to as follows:

> shopping[0]
$R0: String = "Milk"
> costs["Milk"]
$R1: Int? = 1
> shopping.count
$R2: Int = 3
> shopping += [ "Tea" ]
> shopping.count
$R3: Int = 4
> costs.count
$R4: Int = 3
> costs["Tea"] = "String"
error: '@lvalue $T5' is not identical to '(String, Int)'
> costs["Tea"] = 4
> costs.count
$R5: Int = 4

In the previous example, the return type of costs["Milk"] is Int? and not Int. This is an optional type; it represents the possibility of an Int existing. For a dictionary of type T, the return type will be T?. If the value doesn't exist in the dictionary, then the returned value will be nil. Other object-oriented languages, such as Objective-C, C++, Java, and C#, have optional types by default; any object value (or pointer) can be null. By expressing optionality in the type system, Swift can determine whether a value really has to exist or might be nil:

> var cannotBeNil:Int = 1
cannotBeNil: Int = 1
> cannotBeNil = nil
error: type 'Int' does not conform to protocol 'NilLiteralConvertible'
cannotBeNil = nil
> var canBeNil:Int? = 1
canBeNil: Int? = 1
> canBeNil = nil
> canBeNil
$R0: Int? = nil

The value can be tested against nil to find out whether it contains a value and then unpacked with opt!.

As an example, here is how to create and unwrap an optional value:

> var opt: Int? = 1
opt: Int? = 1
> opt == nil
$R1: Bool = false
> opt!
$R2: Int = 1

If a nil value is unpacked, an error occurs:

> opt = nil
> opt!
fatal error: unexpectedly found nil while unwrapping an Optional value
Execution interrupted. Enter Swift code to recover and continue.
Enter LLDB commands to investigate (type :help for assistance.)

Particularly when working with Objective-C based APIs, it is common for values to be declared as optional, although they are always expected to return a value. It is possible to declare such variables as implicitly unwrapped optionals; these variables behave as optional values (they may contain nil), but when the value is accessed, they are automatically unwrapped on demand:

> var implicitlyUnwrappedOptional:Int! = 1
implicitlyUnwrappedOptional: Int! = 1
> implicitlyUnwrappedOptional + 2
3
> implicitlyUnwrappedOptional = nil
> implicitlyUnwrappedOptional + 2
fatal error: unexpectedly found nil while unwrapping an Optional value

Swift has a nil coalescing operator, which is similar to Groovy's ?: operator or C#'s ?? operator. This provides a means to specify a default value if an expression is nil:

> 1 ?? 2
$R0: Int = 1
> nil ?? 2
$R1: Int = 2

The nil coalescing operator can also be used to unwrap an optional value. If the optional value is present, it is unwrapped and returned; if it is missing, then the right-hand side of the expression is returned. Like the shortcut || and && operators, the right-hand side is not evaluated unless necessary:

> costs["Tea"] ?? 0
$R2: Int = 4
> costs["Sugar"] ?? 0
$R3: Int = 0

Conditionally unpacking an optional value is so common that a specific Swift pattern has been created to avoid evaluating the expression twice:

var shopping = [ "Milk", "Eggs", "Coffee", "Tea", ]
var costs = [ "Milk":1, "Eggs":2, "Coffee":3, "Tea":4, ]
var cost = 0
if let cm = costs["Milk"] {
.   cost += cm
. }
> cost
$R0: Int = 1

The if block only executes if the optional value exists. The definition of the constant cm only exists for the body of the if block, and does not exist outside that scope. Furthermore, cm is a non-optional type, so it is guaranteed to not be nil.

To execute an alternative block if the item cannot be found, an else block can be used:

> if let cb = costs["Bread"] {
.  cost += cb
. } else {
.  println("Cannot find any Bread")
. }
Cannot find any Bread

Other boolean expressions can include any expression that conforms to the BooleanType protocol, the equality operators == and !=, the identity operators === and !==, as well as the comparison operators <, <=, >, >=. The is type operator provides a test to see whether an element is of a particular type.

There is a boolean operator specific to Swift, which is the ~= pattern match operator. Despite the name, this isn't anything to do with regular expressions; rather, it's a way of asking whether a pattern matches a particular value. This is used in the implementation of the switch block, which is covered in the next section.

In addition to the if statement, there is a ternary if expression similar to other languages. After a condition, a question mark (?) is used, followed by an expression to be used if the condition is true, then a colon (:) followed by the false expression:

> var i = 17
i: Int = 17
> i % 2 == 0 ? "Even" : "Odd"
$R0: String = "Odd"

In addition to if/else, Swift also has a switch statement, similar to C and Java's switch. However, it differs in two important ways. Firstly, case statements no longer have a default fall-through behavior (so there are no bugs introduced by missing a break statement) and secondly, the value of the case statements can be expressions instead of values, pattern matching on type and range. At the end of the corresponding case, the evaluation jumps to the end of the switch block, unless the fallthrough keyword is used. If no case statements match, the default statement is executed:

> var position = 21
position: Int = 21
> switch position {
.   case 1: println("First")
.   case 2: println("Second")
.   case 3: println("Third")
.   case 4...20: println("\(position)th")
.   case position where (position % 10) == 1:
.     println("\(position)st")
.   case let p where (p % 10) == 2:
.     println("\(p)nd")
.   case let p where (p % 10) == 3:
.     println("\(p)rd")
.   default: println("\(position)th")
. }
21st

In the preceding example, the expression prints out First, Second, or Third if the position is 1, 2, or 3 respectively. For numbers between 4 and 20 (inclusive), it prints out the position with a th ordinal. Otherwise, for numbers that end with 1, it prints st; for numbers that end with 2, it prints nd; and for numbers that end with 3, it prints rd. For all other numbers, it prints th.

The 4...20 range expression in a case statement represents a pattern. If the value of the expression matches that pattern, then the corresponding statements will be executed:

> 4...10 ~= 4
$R0: Bool = true
> 4...10 ~= 21
$R1: Bool = false

There are two range operators in Swift: an inclusive or closed range, and an exclusive or half-open range. The closed range is specified with three dots; 1...12 will give a list of integers between one and twelve. The half-open range is specified with two dots and a less than operator; so 1..<10 will provide integers from 1 to 9 but exclude 10.

The where clause in the switch block allows an arbitrary expression to be evaluated, provided that the pattern matches. These are evaluated in-order, in the sequence they are in the source file. If a where clause evaluates to true, then the corresponding set of statements will be executed.

The let variable syntax can be used to define a constant that refers to the value in the switch block. This local constant can be used in the where clause or the corresponding statements for that specific case. Alternatively, variables can be used from the surrounding scope.

Traversing a dictionary to obtain all of the keys and then subsequently looking up values will result in searching the data structure twice. Instead, both the key and the value can be iterated at the same time using a tuple. A tuple is like a fixed-sized array, but one that allows assigning pairs (or triplets and so on) of values at a time:

> var (a,b) = (1,2)
a: Int = 1
b: Int = 2

Tuples can be used to iterate pairwise over both the keys and values of a dictionary:

> for (item,cost) in costs {
.  println("The \(item) costs \(cost)")
. }
The Coffee costs 3
The Milk costs 1
The Eggs costs 2
The Tea costs 4

Both Array and Dictionary conform to the SequenceType protocol, which allows them to be iterated with a for...in loop. Collections (as well as other objects such as Range) that implement SequenceType have a generate method, which returns a GeneratorType that allows the data to be iterated over. It is possible for custom Swift objects to implement SequenceType to allow them to be used in a for...in loop.

Although the most common use of the for operator in Swift is in a for...in loop, it is also possible to use a more traditional form of for loop. This has an initialization, a condition that is tested at the start of each loop, and a step operation that is evaluated at the end of each loop. Although the parentheses around the for loop are optional, the braces for the block of code are mandatory.

Calculating the sum of integers between 1 and 10 without using the range operator can be done as follows:

> var sum = 0
. for var i=0; i<=10; ++i {
.  sum += i
. }
sum: Int = 55

If multiple variables need to be updated in the for loop, Swift has an expression list that is a set of comma-separated expressions. To step through two sets of variables in a for loop, the following can be used:

> for var i = 0,j = 10; i<=10 && j >= 0; ++i,--j { 
.  println("\(i), \(j)") 
. } 
0, 10
1, 9
…
9, 1
10, 0

The break statement leaves the innermost loop early, and control jumps to the end of the loop. The continue statement takes execution to the top of the innermost loop and the next item.

To break or continue from nested loops, a label can be used. Labels in Swift can only be applied to a loop statement such as while or for. A label is introduced by an identifier and a colon just before the loop statement:

> var deck = [1...13, 1...13, 1...13, 1...13]
> suits: for suit in deck { 
.  for card in suit { 
.   if card == 3 {
.    continue // go to next card in same suit
.   }
.   if card == 5 { 
.     continue suits // go to next suit
.   } 
.   if card == 7 {
.     break // leave card loop
.   }
.   if card == 13 {
.     break suits // leave suit loop
.   }
.  } 
. }     

Swift also supports named arguments, which can either use the name of the variable or can be defined with an external parameter name. To modify the function to support calling with basket and prices as argument names, the following can be done:

> func costOf(basket items:[String], prices costs:[String:Int]) -> Int {
.   var cost = 0
.   for item in items {
.    if let cm = costs[item] {
.     cost += cm
.    }
.   }
.   return cost
. }
> costOf(basket:shopping, prices:costs)
$R1: Int = 10

This example defines external parameter names basket and prices for the function. The function signature is often referred to as costOf(basket:prices:) and is useful when it may not be clear what the arguments are for (particularly if they are for the same type).

A shorthand is available to use the same external name as the parameter name, by prefixing it with a hash (#). These are called shorthand external parameter names:

> func costOf(#items:[String], #costs:[String:Int]) -> Int {
. var cost = 0
. for item in items {
.  if let cm = costs[item] {
.   cost += cm
.  }
. }
. return cost
. }
> costOf(items:shopping, costs:costs)
$R2: Int = 10

Refactoring shorthand external parameter names will lead to API breakage. If it is necessary to change the name internally in a function, convert it from a shorthand name to a separate external and internal parameter name.

Swift functions can have optional arguments by specifying default values in the function definition. When the function is called and an optional argument is missing, the default value for that argument is used.

A default parameter value is specified after the type in the function signature, with an equal sign (=) and then the expression. The expression is re-evaluated each time the function is called without a corresponding value. Default arguments are implicitly named so that the hash (indicating a named argument) is superfluous and will generate warnings.

In the costOf example, instead of passing the value of costs each time, it could be defined with a default parameter as follows:

> func costOf(#items:[String], costs:[String:Int] = costs) -> Int {
.   var cost = 0
.   for item in items {
.    if let cm = costs[item] {
.     cost += cm
.    }
.   }
.   return cost
. }
> costOf(items:shopping)
$R3: Int = 10
> costOf(items:shopping, costs:costs)
$R4: Int = 10

Note that in the first expression, the captured costs variable is bound when the function is defined. If costs is re-assigned at a later stage, then the function will not be updated.

Swift requires that arguments with default values are named, as are arguments that are used in initializers for classes (which are covered in the Classes in Swift section in Chapter 3, Creating an iOS Swift App).

In some cases, this is unnecessary or unhelpful. To disable requiring a named argument for a parameter, the special value underscore (_) can be used:

> func costOf(items:[String], _ costs:[String:Int] = costs) -> Int {
.   var cost = 0
.   for item in items {
.     if let cm = costs[item] {
.       cost += cm
.     }
.   }
.   return cost
. }
> costOf(shopping)
$R0: Int = 10
> costOf(shopping,costs)
$R1: Int = 10

So far, the examples of functions have all returned a single type. What happens if there is more than one return result from a function? In an object-oriented language, the answer is to return a class; however, Swift has tuples, which can be used to return multiple values. The type of a tuple is the type of its constituent parts:

> var pair = (1,2)
pair: (Int, Int) ...

This can be used to return multiple values from the function; instead of just returning one value, it is possible to return a tuple of values.

Separately, it is also possible to take a variable number of arguments. A function can easily take an array of values with [], but Swift provides a mechanism to allow calling with multiple arguments, using variadic functions. The last argument in a function signature can be variadic, which means that it has ellipses after the type. The value can then be used as an array in the function.

Taken together, these two features allow the creation of a minmax function, which returns both the minimum and maximum from a list of integers:

> func minmax(numbers:Int...) -> (Int,Int) {
.   var min = Int.max
.   var max = Int.min
.   for number in numbers {
.    if number < min {
.     min = number
.    }
.    if number > max {
.     max = number
.    }
.   }
.   return(min,max)
. }
> minmax(1,2,3,4)
$R0: (Int, Int) = {
  0 = 1
  1 = 4
}

The numbers:Int... indicates that a variable number of arguments can be passed into the function. Inside the function, it is processed as an ordinary array; in this case, iterating through using a for...in loop.

What if no arguments are passed in? If run on a 64 bit system, then the output will be as follows:

> minmax()
$R1: (Int, Int) = {
  0 = 9223372036854775807
  1 = -9223372036854775808
}

This may not make sense for a minmax function. Instead of returning an error value or a default value, the type system can be used. By making the tuple optional, it is possible to return a nil value if it doesn't exist, or a tuple if it does:

> func minmax(numbers:Int...) -> (Int,Int)? {
.   var min = Int.max
.   var max = Int.min
.   if numbers.count == 0 {
.     return nil
.   } else {
.    for number in numbers {
.     if number < min {
.      min = number
.     } 
.     if number > max {
.      max = number
.     }
.   }
.   return(min,max)
. }
. }
> minmax()
$R2: (Int, Int)? = nil
> mimmax(1,2,3,4)
$R3: (Int, Int)? = (0 = 1, 1 = 3)
> var (minimum,maximum) = minmax(1,2,3,4)!
minimum: Int = 1
maximum: Int = 4

Returning an optional value allows the caller to determine what should happen in cases where the maximum and minimum values are not present.

A tuple is an ordered set of data. The entries in the tuple are ordered, but it can quickly become unclear as to what data is stored, particularly if they are of the same type. In the minmax tuple, it is unclear which value is the minimum and which is the maximum, and this can lead to subtle programming errors later on.

A structure is like a tuple, but with named values. This allows members to be accessed by name instead of by position, leading to fewer errors and greater transparency. Named values can be added to tuples as well. In essence, tuples with named values are anonymous structures.

A struct is defined with the keyword struct and has variables or values in the body:

> struct MinMax {
.   var min:Int
.   var max:Int
. }

This defines a MinMax type, which can be used in place of any of the types seen so far. It can be used in the minmax function to return a struct instead of a tuple:

> func minmax(numbers:Int…) -> MinMax? {
.   var minmax = MinMax(min:Int.max, max:Int.min)
.   if numbers.count == 0 {
.     return nil
.   } else {
.     for number in numbers {
.       if number < minmax.min {
.         minmax.min = number
.        }
.       if number > minmax.max {
.         minmax.max = number
.       }
.     }
.     return minmax
.   }
. }

The struct is initialized with a type constructor; if MinMax() is used, then the default values for each of the structure members are used (based on the structure definition), but these defaults can be overridden explicitly if desired, with MinMax(min:-10,max:11). For example, if the MinMax struct is defined as struct MinMax { var min:Int = Int.max; var max:Int = Int.min }, then MinMax() would return a structure with the appropriate maximum and minimum values filled in.

Swift also has classes; these are covered in the Swift classes section in the next chapter.

Save the following as hello.swift:

#!/usr/bin/env xcrun swift
println("Hello World")

After saving, make the file executable by running chmod a+x hello.swift. The program can then be run by typing ./hello.swift, and the traditional greeting will be seen:

Hello World

Arguments can be passed in from the command line and interrogated in the process using the Process class through the arguments constant. As with other Unix commands, the first element (0) is the name of the process executable; the arguments passed on in the command line start from one (1).

The program can be terminated using the exit function; however, this is defined in the Foundation framework and so it needs to be imported in order to call this function. Modules in Swift correspond to Frameworks in Objective-C and give access to all functions defined as public API in the module. The syntax to import all elements from a module is import module, although it's also possible to import a single function using import func module.functionName.

A Swift program to print arguments in uppercase can be implemented as follows:

#!/usr/bin/env xcrun swift
import Foundation
let args = Process.arguments[1..<countElements(Process.arguments)]
for arg in args {
  println("\(arg.uppercaseString)")
}
exit(0)

Running this with hello world results in the following:

$ ./upper.swift hello world
HELLO
WORLD

Conventionally, the entry point to Swift programs is via a script called main.swift. If starting a Swift-based command-line application project in Xcode, a main.swift file will be created automatically. Scripts do not need to have a .swift extension. For instance, the previous example could be called upper and it would still work.

While interpreted Swift scripts are useful for experimenting and writing, each time the script is started, it is interpreted using the Swift command-line tool and then executed. For simple scripts (such as converting arguments to upper case), this can be a large proportion of the script's execution time.

To compile a Swift script into a native executable, use the swiftc command with the -o output flag to specify a file to write to. This will then generate an executable that does exactly the same as the interpreted script, only much faster. The time command can be used to compare the running time of the interpreted and compiled versions:

$ time ./upper.swift hello world # Interpreted
HELLO
WORLD
real  0m0.145s
$ xcrun swiftc -o upper upper.swift     # Compile step
$ time ./upper hello world       # Compiled
HELLO
WORLD
real  0m0.012s

Of course, the numbers will vary and the initial step only happens once, but startup is very lightweight in Swift. The numbers mentioned earlier are not meant to be taken in magnitude but rather as relative to each other.

The compile step can also be used to link together many individual Swift files into one executable, which helps create a more organized project; Xcode will encourage having multiple Swift files as well.