Book Image

Beginning Swift

By : Rob Kerr, Kåre Morstøl
Book Image

Beginning Swift

By: Rob Kerr, Kåre Morstøl

Overview of this book

Take your first foray into programming for Apple devices with Swift.Swift is fundamentally different from Objective-C, as it is a protocol-oriented language. While you can still write normal object-oriented code in Swift, it requires a new way of thinking to take advantage of its powerful features and a solid understanding of the basics to become productive.
Table of Contents (10 chapters)

Swift Data Types

Like most programming languages, Swift includes a full complement of built-in data types that store numbers, characters, strings, and Boolean values.


In the previous section, we covered the use of Swift optionals, and worked through several examples declaring an Int variable as optional and non-optional. Keep in mind that any Swift variable, of any type, can be declared as an optional.

Numeric Data Types

Like most programming languages, Swift provides built-in numeric data types that represent either integer or floating-point values.

Int on 64-Bit Versus 32-Bit Platforms

While it's likely you'll develop Swift applications exclusively on 64-bit platforms, it's important to know that Swift is available on both 32-bit and 64-bit platforms. When using a generic integer numeric type ( Int or UInt), the generic type will be mapped to an underlying, specific equivalent that matches the current platform's word size. For example, on a 64-bit platform, Int is mapped to Int64; on a 32-bit platform, the same Int type is mapped to an Int32.

Built-In Numeric Data Types

The following table summarizes the available Swift numeric data types:


Min value

Max value








-2.1 x 109

2.1 x 109


-9.2 x 1018

9.2 x 1018









4.3 x 109



1.8 x 1019


-1.8 x 10308

1.8 x 10308


-3.4 x 1038

3.4 x 1038

Choosing the Appropriate Numeric Data Type

Conceptually, a UInt64 variable will consume four times more RAM than a UInt8 variable, so you may ask, " Should I tune my variables by selecting the smallest number of bits needed to meet requirements?"

While it may seem intuitive to select the numeric type that uses the least RAM to store the variable's expected range of values, it's usually preferable to use the generic integer types (for example, Int when declaring integers and Double when declaring floating-point numbers).


This is a reference from The Swift Programming Language (Swift 4): " Unless you need to work with a specific size of integer, always use Int for integer values in your code. This aids code consistency and interoperability." Visit for the official documentation.

Declaring and Assigning Integer Variables

Integer values may be instantiated using base 10 (decimal), base 2 (binary), base 8 (octal), or base 16 (hexadecimal) literal values, or by assigning another Int variable of the same type to the new variable.

For example, assigning the number 100 to a new Int variable holding a duration in minutes can be done in any of the following ways:

let minutes = 100         // decimal
let minutes = 0b1100100   // binary
let minutes = 0o144       // octal
let minutes = 0x64        // hexadecimal

Declaring and Assigning Floating Point Numbers

Floating-point numbers are represented by either Float or Double data types. In general, you should use Double—and employ Float only when specific circumstances require using the smaller, 32-bit numeric variable.

Declaring and assigning value to floating-point variables follows the same syntax rules as with integer variables. For example, the following statement creates a new Double variable interestRate, and assigns an initial value to it:

var interestRate = 5.34

Numeric Literal Grouping

When assigning constant values to numeric types, Swift provides a handy format to make code more readable: the underscore character is ignored when parsing numeric literals.

This feature is most commonly used to provide groupings of thousands in a large integer or floating-point assignments, but actually can be used to provide any grouping separation that makes code more readable. For example, the following statements all assign the value 100,000 to the variable minutes:

var minutes = 100000
var minutes = 100_000
var minutes = 10_00_00
var minutes = 0b110_000110_101000_00

Using the underscore for readability can also be used for floating-point literal values. For example, the following statements are equivalent:

var balance = 10000.44556
var balance = 10_000.44_556

Numeric Type Conversions

Like many fully compiled languages, Swift is a strongly typed language, and requires explicit type conversions (or casts) when assigning the value from one variable type to a variable of a different type.

Many new Swift programmers find that Swift is even stricter than languages they've used before. In many programming languages, the compiler will implicitly convert between data types during an assignment so long as the value contained within the variable being assigned (on the right of the equals sign) could not overflow the variable being assigned to (on the left of the equals sign).

In other words, in many languages, the following code would be legal, since an Int8 is known to always fit into an Int16 without a numeric overflow:

Int8 smallNumber = 3;
Int16 mediumNumber = smallNumber;

However, this equivalent code in Swift would result in a compile-time error:

var smallNumber: Int8 = 3
var mediumNumber: Int16 = smallNumber

This code would generate the following error:

error: cannot convert value of type 'Int8' to specified type 'Int16'

In Swift, it's always the programmer's responsibility to ensure that assignments have the same data type on the left and right of the assignment operator (that is, the equals sign). The following code corrects the compile-time error:

var smallNumber: Int8 = 100
var mediumNumber: Int16 = Int16(smallNumber)


This requirement for explicit type assignment is one reason why most Swift programming uses the generic numeric variables Int and Double, except when specific usage requires tuning for numeric range or memory storage size.

Using Numeric Types

Now, let's see how to use various numeric variable types by following these steps:

  1. Launch Xcode as before, and create a new playground named Topic B Using Numeric Types.playground.

  2. Add the following code to the playground to create three Int variables, using binary, base10, and base16 literal notation, respectively:

    var base2 = 0b101010
    var base10 = 42
    var hex = 0x2A
  3. Now add the following three corresponding lines to print the data type and value for each of the variables you just created.

    print("Printing \(type(of: base2)): \(base2)")
    print("Printing \(type(of: base10)): \(base10)")
    print("Printing \(type(of: hex)): \(hex)")

    Examining the output, note that the three variables all have the same data type (Int) and same value (42 in base 10).

  4. Add the following lines of code to create two more variables, and to print the types and values for each:

    var scientific = 4.2E+7
    let double = 4.99993288828
    print("Printing \(type(of: scientific)): \(scientific)")
    print("Printing \(type(of: double)): \(double)")

    Note that both variables were created as Double types—even though the value of the first is actually an Integer. Swift's inference system doesn't always look at the actual value. In this case, the presence of scientific notation in the literal value caused Swift to assume the value should be a Double.

  5. Now add the following lines to cast and round the variable named double to an Int:

    var castToInt = Int(double)
    var roundToInt = Int(double.rounded())
    print("Printing \(type(of: castToInt)): \(castToInt)")
    print("Printing \(type(of: roundToInt)): \(roundToInt)")

    As you probably expected, the castToInt discarded the fractional value of the original double variable. For the roundToInt variable, we called the .rounded() function on the variable double, and then cast that value. Since 4.999 was rounded up to 5 before being cast, the Int contains the rounded value.

  6. Finally, add the following lines to create a very large unsigned integer and then print its type and value:

    var bigUnsignedNumber:UInt64 = 18_000_000_000_000_000_000
    print("Printing \(type(of: bigUnsignedNumber)): \(bigUnsignedNumber)")

    This code works as expected—printing an integer with 20 digits (the underscore is added to help count how many digits there are).

    Note that in this case, we specified UInt64 should be the data type for this variable. Had we not made the type explicit, Swift's type inference rules would have assigned the smaller Int data type to the variable, and it would have overflowed.

Again, keep in mind the inference engine examines the format of a constant perhaps more than the value of the numeric value being assigned. You should rely on the inference engine by default, but keep in mind you may sometimes need to be explicit when you know more about how a variable will be used than Swift can infer.


In Swift, the Boolean data type is Bool, and stores a value of true or false. As with other data types, in the case that a Bool value is not yet known, a Bool can be declared as optional, for example, Bool?.

For example, the following code declares a Boolean in Swift, and then changes its value:

var isChecked = false
isChecked = true

Testing for the value of a Bool value is similar to how we do it in other C-inspired languages, for example:

if isChecked {
   // statements to execute if isChecked is true
if isChecked == true {
   // statements to execute if isChecked is true
if !isChecked {
   // statements to execute if isChecked is false


The Character data type in Swift is an extended grapheme cluster.

What does that mean?

An extended grapheme cluster is an ordered sequence of one or more Unicode scalars (that is, values) that, when taken together, produce a human-readable character.

Most important to understand is that, unlike ASCII or ANSI character representations many programmers have worked with before, a Character in Swift may be made of more than one Unicode value.

In Swift 4, the underlying complexities of Unicode, scalar values, and extended grapheme clusters are largely managed for you, but as you begin to work natively with Unicode characters and strings, bear in mind that the Swift Character/String architecture was developed from the ground up around Unicode character representation—not ANSI/ASCII as many other languages were.

Assigning a Character

The following are two examples creating new Character variables, and assigning literal values:

let ch1:Character = "A"
let ch2:Character = "😎"

Note the following regarding this assignment:

  • In Swift, a Character literal is delimited by a double quote, rather than the single quote that's common in most C-inspired languages.

  • Because the Swift compiler's type inference rules will assume double quotes around a literal imply a string variable, the above ch1 assignment must explicitly declare the variables as Character type—otherwise the Swift compiler will create ch1 as a string.

Constructing a Character Literal

To construct a Character type using Unicode values, you can assign an escape sequence, or use the UnicodeScalar struct to create a Character using numeric Unicode values as input.

The following line of code creates a UnicodeScalar from the value 65 (the ASCII value for the English letter A), and then assigns it to the immutable variable ch1:

let ch1 = Character(UnicodeScalar(65))

In this case, there is no ambiguity with regards to double quotation marks, so it's not necessary to explicitly assign the Character type during this assignment.

It's also common to construct a Character using a UnicodeScalar escape sequence within double quotation marks. The following creates a character variable containing an emoji character represented by the UnicodeScalar 1F601:

let ch3 = "\u{1F601}"  // sets ch3 to "😁"

While Unicode scalars are conceptually similar to ASCII/ANSI value encoding, Swift Characters may be made of more than one numeric value, while ASCII and ANSI use only one numeric value to represent each character.

For example, an accented Western letter is expressed by providing a UnicodeScalar containing two character values.

We can construct the Unicode representation of an accented e character as follows:

let ch4 = "e\u{301}"   // é

The expression on the right of the assignment contains the literal letter e, followed by the escaped value for the accent modifier (301). The Swift compiler combines these two elements into a single extended grapheme cluster.


Strings in Swift are very similar to strings in other programming languages. As string handling is so central to any application development project, we'll dedicate an entire subsequent lesson to Swift's powerful string handling capabilities. In this section, we'll discuss the basics for declaring and using a string.

Fundamentally, strings are arrays of the Character types, supporting the familiar assignment operator (=), substrings, concatenation, and C-inspired escape characters.

Instantiating a String

Instantiating a string variable is highly intuitive. The following statements create string variables:

let macCharacters = "⌘⌃⌥⇧ ⏎⌫⇪⎋⇥"
let emoji = "😎😂🎃🐳🍎😜😆"

String Concatenation

As in many languages, Swift strings can be concatenated using the plus (+) operator:

let alphaMac = alphabet + macCharacters

String also supports the unary addition operator:

alphabet += macCharacters

Extracting Characters

One difference between Swift strings and strings in many languages is how individual elements of strings are accessed. Specifically, the following syntax with Swift strings is illegal:

let ch = alphabet[4]
error: 'subscript' is unavailable: cannot subscript String with an Int, see the documentation comment for discussion

In Swift, the input to the subscript operator (that is, what's between the [] characters) is expected to be of type String.Index, not Int.

In practice, you will construct an Index, then pass the index to the substring operator, for example:

let idx = alphabet.index(alphabet.startIndex, offsetBy: 4)
let ch = alphabet[idx]  // ch is assigned the character "E"

String Length

Obtaining the length of string is quite easy—simply call the count property of a string:

let alphabetLength = alphabet.count  // 26

We have now reached the end of this section. Here, we worked with the different data types in Swift, specifically numeric, Boolean, character, and string data types.

Activity: Data Type Summary

Now that you've learned about the various data types available with Swift, let's put this knowledge into practice by using various types together, and also using the Apple Foundation framework.

Use an Xcode playground to practice various data types. You'll be using numeric data types, formatting them as strings, and using string interpolation to print string values from various data types.

  1. Launch Xcode as before, and create a new playground named Data Type Summary.playground.

  2. Add the following code to the playground to create an immutable Double with an initial value:

    let dVal = 4.9876
  3. Next, create a Boolean mutable variable with an initial value of true, and another variable set to the Double variable after rounding to a whole number:

    var iValRounded = true
    var iVal = Int(dVal.rounded())
  4. Next, we're going to use a class from Foundation to create a string representation of the Double value, rounded to two digits. If you're not familiar with NumberFormatter, don't worry. This is just one of the many utility classes Apple provides in its expansive SDK for macOS and iOS:

    var formatDigits = 2
    let nf = NumberFormatter()
    nf.numberStyle = .decimal
    nf.maximumFractionDigits = formatDigits
    let formattedDouble = nf.string(from: NSNumber(value: dVal)) ?? "#Err"

    Because NumberFormatter.string returns an optional, we need either to check it (with if/let, or as here, provide a default value ("#Err") in case the function does return nil.

  5. Now add the following line to print a statement about the values we've created:

    print("The original number was \(formattedDouble) (rounded to \(formatDigits) decimal places), while the value \(iValRounded ? "rounded" : "unrounded") to Integer is \(iVal).")

    The output of this code is as follows:

    The original number was 4.99 (rounded to 2 decimal places), while the value rounded to Integer is 5.
  6. Finally, add the following lines to change the rounding strategy, and print a sentence about the result of the new string conversions:

    formatDigits = 0
    nf.maximumFractionDigits = formatDigits
    formattedDouble = nf.string(from: NSNumber(value: dVal)) ?? "#Err"
    iValRounded = false
    iVal = Int(dVal)
    print("The original number was \(formattedDouble) (rounded to \(formatDigits) decimal places), while the value \(iValRounded ? "rounded" : "unrounded") to Integer is \(iVal).")

    The output of this second sentence is as follows:

    The original number was 5 (rounded to 0 decimal places), while the value unrounded to Integer is 4.