kotlin-for-python-developers

Kotlin for Python developers

By Aasmund Eldhuset, Software Engineer at Khan Academy. Published on November 29, 2018.
This document is not a part of Khan Academy’s official product offering, but rather an internal resource that we’re providing “as is” for the benefit of the programming community. If you find any errors, please submit an issue or a pull request.


Kotlin is a compiled, statically typed language, which might provide some initial hurdles for people who are used to the interpreted, dynamically typed Python. This document aims to explain a substantial portion of Kotlin’s syntax and concepts in terms of how they compare to corresponding concepts in Python.

Kotlin can be compiled for several different platforms. In this document, we assume that the target platform is the Java virtual machine, which grants some extra capabilities - in particular, your code will be compiled to Java bytecode and will therefore be interoperable with the large ecosystem of Java libraries.

Even if you don’t know Python, this document should hopefully be a useful introduction to Kotlin, in particular if you are used to other dynamically typed languages. However, if you’re coming from a Java background, you’re probably better off diving directly into the excellent official docs (from which this doc has drawn a lot of inspiration). To some extent, you can try to write Java code and look stuff up whenever what you’re trying to do doesn’t work - and some IDEs can even automatically convert Java code to Kotlin.

Contents

Hello World

Let’s get straight to the point - type this into a file with the extension .kt:

fun main(args: Array<String>) {
    println("Hello World!")
}

Only imports and declarations can exist at the top level of a Kotlin file. Therefore, “running” an individual file only makes sense if it contains an entry point, which must be a function called main with one argument called args of the type “array of strings”. args will contain the command-line arguments that the program is invoked with, similarly to sys.argv in Python; it can be omitted if your program does not need to accept command-line arguments and you are using Kotlin 1.3:

fun main() {
    println("Hello World!")
}

The function body is delimited by curly braces - indentation is generally not significant in Kotlin, but you should of course indent your code properly for the benefit of human readers.

Comments are initiated with // and last until the end of the line. Block comments start with /* and end with */.

Like in Python, statements may be terminated by a semicolon, but it’s discouraged. There is no line continuation character; instead, a line is automatically joined with one or more of the subsequent lines if that’s the only way to make the code parse correctly. In practice, that means that a statement continues on the next line if we’re inside an open parenthesis (like in Python), or if the line ends with a “dangling operator” (unlike in Python) or the following line doesn’t parse unless it’s joined to the previous one (also unlike in Python). Note that this is pretty much the opposite of JavaScript, which generally will keep joining lines as long as the resulting code still parses. Thus, the following is two expressions in Kotlin and in Python (because + can be unary, so the second line parses on its own), but one in JavaScript:

1 + 2
+ 3

This is one expression in both Kotlin (because the first line doesn’t parse on its own) and JavaScript, and doesn’t parse in Python:

1 + 2 +
3

So is the following. The difference between + and . is that + can be a unary operator, but . can’t, so the only way to get the second line to parse is to join it with the preceding line:

x.foo()
 .bar()

This is one expression in all three languages:

(1 + 2
 + 3)

Don’t split lines if the resulting two lines are also grammatically valid on their own (even if it results in a compilation error that is not directly related to the grammar of Kotlin). The following does not actually return the result of foo() - it returns a special value called Unit, which we’ll cover later, and foo() is never called.

return    // Empty return statement
    foo() // Separate, unreachable statement

Compiling and running

The author strongly recommends that you use an IDE with Kotlin support, as the static typing allows an IDE to do reliable navigation and code completion. I recommend IntelliJ IDEA, which is built by the same company that created Kotlin. The Community Edition is free; see instructions for getting started (it comes bundled with Kotlin, and you can run your program from the IDE).

If you insist on using a plain editor and the command line, see these instructions instead. In short, you need to compile your Kotlin code before running it. Assuming that your Kotlin file is called program.kt:

kotlinc program.kt -include-runtime -d program.jar

By default, Kotlin compiles down to Java (so you have the entire Java Standard Library available to you, and interacting with Java libraries is a breeze), so you now have a Java Archive (program.jar) which includes the Java libraries that are necessary to support the Kotlin features (thanks to -include-runtime), and you can run it using an out-of-the-box Java runtime:

java -jar program.jar

Declaring variables

Every variable must be declared. Any attempt to use a variable that hasn’t been declared yet is a syntax error; thus, you are protected from accidentally assigning to a misspelled variable. The declaration also decides what kind of data you are allowed to store in the variable.

Local variables are typically declared and initialized at the same time, in which case the type of the variable is inferred to be the type of the expression you initialize it with:

var number = 42
var message = "Hello"

We now have a local variable number whose value is 42 and whose type is Int (because that’s the type of the literal 42), and another local variable message whose value is "Hello" and whose type is String. Subsequent usages of the variable must use only the name, not var:

number = 10
number += 7
println(number)
println(message + " there")

However, you cannot change the type of a variable: number can only ever refer to Int values, and message can only ever refer to String values, so both number = "Test" and message = 3 are illegal and will produce syntax errors.

Read-only variables

Frequently, you’ll find that during the lifetime of your variable, it only ever needs to refer to one object. Then, you can declare it with val (for “value”) instead:

val message = "Hello"
val number = 42

The terminology is that var declares a mutable variable, and that val declares a read-only or assign-once variable - so both kinds are called variables.

Note that a read-only variable is not a constant per se: it can be initialized with the value of a variable (so its value doesn’t need to be known at compile-time), and if it is declared inside a construct that is repeatedly invoked (such as a function or a loop), it can take on a different value on each invocation. Also, while the read-only variable may not be reassigned while it is in scope, it can still refer to an object which is in itself mutable (such as a list).

Constants

If you have a value that is truly constant, and the value is a string or a primitive type (see below) that is known at compile-time, you can declare an actual constant instead. You can only do this at the top level of a file or inside an object declaration (but not inside a class declaration):

const val x = 2

Specifying the type explicitly

If you really want to, you can both initialize and specify the type on the same line. This is mostly useful if you’re dealing with a class hierarchy (more on that later) and you want the variable type to be a base type of the value’s class:

val characters: CharSequence = "abc"

In this doc, we’ll sometimes specify the type unnecessarily, in order to highlight what type is produced by an expression. (Also, a good IDE will be able to show you the resulting type.)

For completeness: it is also possible (but discouraged) to split the declaration and the initial assignment, and even to initialize in multiple places based on some condition. You can only read the variable at a point where the compiler can prove that every possible execution path will have initialized it. If you’re creating a read-only variable in this way, you must also ensure that every possible execution path assigns to it exactly once.

val x: String
x = 3

Scopes and naming

A variable only exists inside the scope (curly-brace-enclosed block of code; more on that later) in which it has been declared - so a variable that’s declared inside a loop only exists in that loop; you can’t check its final value after the loop. Variables can be redeclared inside nested scopes - so if there’s a parameter x to a function and you create a loop and declare an x inside that loop, the x inside the loop is a different variable than the parameter x.

Variable names should use lowerCamelCase instead of snake_case.

In general, identifiers may consist of letters, digits, and underscores, and may not begin with a digit. However, if you are writing code that e.g. autogenerates JSON based on identifiers and you want the JSON key to be a string that does not conform to these rules or that collides with a keyword, you can enclose it in backticks: `I can't believe this is not an error!` is a valid identifier.

Primitive data types and their limitations

The primitive data types are the most fundamental types in Kotlin; all other types are built up of these types and arrays thereof. Their representation is very efficient (both in terms of memory and CPU time), as they map to small byte groups that are directly manipulatable by the CPU.

Integer types

Integer types in Kotlin have a limited size, as opposed to the arbitrarily large integers in Python. The limit depends on the type, which decides how many bits the number occupies in memory:

Type Bits Min value Max value
Long 64 -9223372036854775808 9223372036854775807
Int 32 -2147483648 2147483647
Short 16 -32768 32767
Byte 8 -128 127

Bytes are -128 through 127 due to Kotlin inheriting a bad design decision from Java. In order to get a traditional byte value between 0 and 255, keep the value as-is if it is positive, and add 256 if it is negative (so -128 is really 128, and -1 is really 255). See the section on extension functions for a neat workaround for this.

An integer literal has the type Int if its value fits in an Int, or Long otherwise. Long literals should be suffixed by L for clarity, which will also let you make a Long with a “small” value. There are no literal suffixes for Short or Byte, so such values need an explicit type declaration or the use of an explicit conversion function.

val anInt = 3
val anotherInt = 2147483647
val aLong = 2147483648
val aBetterLong = 2147483649L
val aSmallLong = 3L
val aShort: Short = 32767
val anotherShort = 1024.toShort()
val aByte: Byte = 65
val anotherByte = -32.toByte()

Beware that dividing an integer by an integer produces an integer (like in Python 2, but unlike Python 3). If you want a floating-point result, at least one of the operands needs to be a floating-point number (and recall that like in most languages, floating-point operations are generally imprecise):

println(7 / 3)            // Prints 2
println(7 / 3.0)          // Prints 2.3333333333333335
val x = 3
println(7 / x)            // Prints 2
println(7 / x.toDouble()) // Prints 2.3333333333333335

Whenever you use an arithmetic operator on two integers of the same type (or when you use a unary operator like negation), there is no automatic “upgrading” if the result doesn’t fit in the type of the operands! Try this:

val mostPositive = 2147483647
val mostNegative = -2147483648
println(mostPositive + 1)
println(-mostNegative)

Both of these print -2147483648, because only the lower 32 bits of the “real” result are stored.

When you use an arithmetic operator on two integers of different types, the result is “upgraded” to the widest type. Note that the result might still overflow.

In short: think carefully through your declarations of integers, and be absolutely certain that the value will never ever need to be larger than the limits of the type! If you need an integer of unlimited size, use the non-primitive type BigInteger.

Floating-point and other types

Type Bits Notes
Double 64 16-17 significant digits (same as float in Python)
Float 32 6-7 significant digits
Char 16 UTF-16 code unit (see the section on strings - in most cases, this is one Unicode character, but it might be just one half of a Unicode character)
Boolean 8 true or false

Floating-point numbers act similarly to in Python, but they come in two types, depending on how many digits you need. If you need larger precision, or to work with monetary amounts (or other situations where you must have exact results), use the non-primitive type BigDecimal.

Strings

Unicode correctness can be onerous in Python 2, since the “default” string type str is really just a byte array, while unicode is actually a sequence of code units (see below) - and whether the code units are 16 or 32 bits wide depends on how your Python distribution was built. In Kotlin, there’s no such confusion: String, which is what you get when you make a string literal (which you can only do with double quotes), is an immutable sequence of UTF-16 code units. ByteArray is a fixed-size (but otherwise mutable) byte array (and String can specifically not be used as a byte array).

A UTF-16 code unit is a 16-byte unsigned integral value that represents either one Unicode code point (character code) or must be combined with another code unit to form a code unit. If this makes no sense, I strongly recommend Joel Spolsky’s excellent essay on Unicode and its encodings. For most Western scripts, including English, all code points fit inside one code unit, so it’s tempting to think of a code unit as a character - but that will lead astray once your code encounters non-Western scripts. A single UTF-16 code unit can be represented with single quotes, and has the type Char:

val c = 'x' // Char
val message = "Hello" // String
val m = message[0] // Char

Thus, single quotes can not be used to form string literals.

Given a string s, you can get a ByteArray with the UTF-8 encoding of the string by calling s.toByteArray(), or you can specify another encoding, e.g. s.toByteArray(Charsets.US_ASCII) - just like encode() in Python. Given a byte array b that contains a UTF-8-encoded string, you can get a String by calling String(b); if you’ve got a different encoding, use e.g. String(b, Charsets.US_ASCII), just like decode() in Python. You can also call e.g. b.toString(Charsets.US_ASCII), but do not call b.toString() without parameters (this will just print an internal reference to the byte array).

You can do string interpolation with $, and use curly braces for expressions:

val name = "Anne"
val yearOfBirth = 1985
val yearNow = 2018
val message = "$name is ${yearNow - yearOfBirth} years old"

If you want a literal $, you need to escape it: \$. Escaping generally works the same way as in Python, with a similar set of standard escape sequences.

Conditionals

if/else

if/else works the same way as in Python, but it’s else if instead of elif, the conditions are enclosed in parentheses, and the bodies are enclosed in curly braces:

int age = 42
if (age < 10) {
    println("You're too young to watch this movie")
} else if (age < 13) {
    println("You can watch this movie with a parent")
} else {
    println("You can watch this movie")
}

The curly braces around a body can be omitted if the body is a oneliner. This is discouraged unless the body goes on the same line as the condition, because it makes it easy to make this mistake, especially when one is used to Python:

if (age < 10)
    println("You're too young to watch this movie")
    println("You should go home") // Mistake - this is not a part of the if body!

Without the curly braces, only the first line is a part of the body. Indentation in Kotlin matters only for human readers, so the second print is outside the if and will always be executed.

An if/else statement is also an expression, meaning that a ternary conditional (which looks like result = true_body if condition else false_body in Python) looks like this in Kotlin:

val result = if (condition) trueBody else falseBody

When using if/else as an expression, the else part is mandatory (but there can also be else if parts). If the body that ends up being evaluated contains more than one line, it’s the result of the last line that becomes the result of the if/else.

Comparisons

Structural equality comparisons are done with == and !=, like in Python, but it’s up to each class to define what that means, by overriding equals() (which will be called on the left operand with the right operand as the parameter) and hashCode(). Most built-in collection types implement deep equality checks for these operators and functions. Reference comparisons - checking if two variables refer to the same object (the same as is in Python) - are done with === and !==.

Boolean expressions are formed with && for logical AND, || for logical OR, and ! for logical NOT. As in Python, && and || are short-circuiting: they only evaluate the right-hand side if it’s necessary to determine the outcome. Beware that the keywords and and or also exist, but they only perform bitwise operations on integral values, and they do not short-circuit.

There are no automatic conversions to boolean and thus no concept of truthy and falsy: checks for zero, empty, or null must be done explicitly with == or !=. Most collection types have an isEmpty() and an isNotEmpty() function.

when

We’re not going to cover the when expression in depth here since it doesn’t have a close equivalent in Python, but check it out - it’s pretty nifty, as it lets you compare one expression against many kinds of expressions in a very compact way (but it’s not a full functional-programming-style pattern matcher). For example:

val x = 42
when (x) {
    0 -> println("zero")
    in 1..9 -> println("single digit")
    else -> println("multiple digits")
}

Collections

Arrays in Kotlin have a constant length, so one normally uses lists, which are similar to the ones in Python. What’s called a dict in Python is called a map in Kotlin (not to be confused with the function map()). List, Map, and Set are all interfaces which are implemented by many different classes. In most situations, a standard array-backed list or hash-based map or set will do, and you can easily make those like this:

val strings = listOf("Anne", "Karen", "Peter") // List<String>
val map = mapOf("a" to 1, "b" to 2, "c" to 3)  // Map<String, Int>
val set = setOf("a", "b", "c")                 // Set<String>

(Note that to is an infix function that creates a Pair containing a key and a value, from which the map is constructed.) The resulting collections are immutable - you can neither change their size nor replace their elements - however, the elements themselves may still be mutable objects. For mutable collections, do this:

val strings = mutableListOf("Anne", "Karen", "Peter")
val map = mutableMapOf("a" to 1, "b" to 2, "c" to 3)
val set = mutableSetOf("a", "b", "c")

You can get the size/length of a collection c with c.size (except for string objects, where you for legacy Java reasons must use s.length instead).

Unfortunately, if you want an empty collection, you need to either declare the resulting collection type explicitly, or supply the element type(s) to the function that constructs the collection:

val noInts: List<Int> = listOf()
val noStrings = listOf<String>()
val emptyMap = mapOf<String, Int>()

The types inside the angle brackets are called generic type parameters, which we will cover later. In short, it’s a useful technique to make a class that is tied to another class (such as a container class, which is tied to its element class) applicable to many different classes.

If you really really need a mixed-type collection, you can use the element type Any - but you’ll need typecasting to get the elements back to their proper type again, so if what you want is a multiple-value return from a function, please use the per-element-typed Pair or Triple instead. If you need four or more elements, consider making a data class for the return type instead (which you should ideally do for two or three elements as well, especially if it’s a public function, since it gives you proper names for the elements) - it’s very easy and usually a oneliner.

Loops

for

Kotlin’s loops are similar to Python’s. for iterates over anything that is iterable (anything that has an iterator() function that provides an Iterator object), or anything that is itself an iterator:

val names = listOf("Anne", "Peter", "Jeff")
for (name in names) {
    println(name)
}

Note that a for loop always implicitly declares a new read-only variable (in this example, name) - if the outer scope already contains a variable with the same name, it will be shadowed by the unrelated loop variable. For the same reason, the final value of the loop variable is not accessible after the loop.

You can also create a range with the .. operator - but beware that unlike Python’s range(), it includes its endpoint:

for (x in 0..10) println(x) // Prints 0 through 10 (inclusive)

If you want to exclude the last value, use until:

for (x in 0 until 10) println(x) // Prints 0 through 9

You can control the increment with step:

for (x in 0 until 10 step 2) println(x) // Prints 0, 2, 4, 6, 8

The step value must be positive. If you need to count downwards, use the inclusive downTo:

for (x in 10 downTo 0 step 2) println(x) // Prints 10, 8, 6, 4, 2, 0

Any of the expressions to the right of in in the loops above can also be used outside of loops in order to generate ranges (one type of iterables - this is similar to xrange() in Python 2 and range() in Python 3), which can be iterated over later or turned into lists:

val numbers = (0..9).toList()

If you need to know the index of the current element when you’re iterating through something, you can use withIndex(), which corresponds to enumerate(). It produces a sequence of objects that have got two properties (the index and the value) and two specially-named accessor functions called component1() and component2(); Kotlin lets you destructure such an object into a declaration:

for ((index, value) in names.withIndex()) {
    println("$index: $value")
}

while

The while loop is similar to Python (but keep in mind that the condition must be an actual boolean expression, as there’s no concept of truthy or falsy values).

val x = 0
while (x < 10) {
    println(x)
    x++ // Same as x += 1
}

The loop variable(s), if any, must be declared outside of the while loop, and are therefore available for inspection afterwards, at which point they will contain the value(s) that made the loop condition false.

continue and break

A plain continue or break works the same way as in Python: continue skips to the next iteration of the innermost containing loop, and break stops the loop. However, you can also label your loops and reference the label in the continue or break statement in order to indicate which loop you want to affect. A label is an identifier followed by @, e.g. outer@ (possibly followed by a space). For example, to generate primes:

outer@ for (n in 2..100) {
    for (d in 2 until n) {
        if (n % d == 0) continue@outer
    }
    println("$n is prime")
}

Note that there must be no space between continue/break and @.

Functions

Declaration

Functions are declared with the fun keyword. For the parameters, you must declare not only their names, but also their types, and you must declare the type of the value the function is intending to return. The body of the function is usually a block, which is enclosed in curly braces:

fun happyBirthday(name: String, age: Int): String {
    return "Happy ${age}th birthday, $name!"
}

Here, name must be a string, age must be an integer, and the function must return a string. However, you can also make a oneliner function, where the body simply is the expression whose result is to be returned. In that case, the return type is inferred, and an equals sign is used to indicate that it’s a oneliner:

fun square(number: Int) = number * number

(Note that there is no ** operator; non-square exponentiation should be done via Math.pow().)

Function names should use lowerCamelCase instead of snake_case.

Calling

Functions are called the same way as in Python:

val greeting = happyBirthday("Anne", 32)

If you don’t care about the return value, you don’t need to assign it to anything.

Returning

As opposed to Python, omitting return at the end of a function does not implicitly return null; if you want to return null, you must do so with return null. If a function never needs to return anything, the function should have the return type Unit (or not declare a return type at all, in which case the return type defaults to Unit). In such a function, you may either have no return statement at all, or say just return. Unit is both a singleton object (which None in Python also happens to be) and the type of that object, and it represents “this function never returns any information” (rather than “this function sometimes returns information, but this time, it didn’t”, which is more or less the semantics of returning null).

Overloading

In Python, function names must be unique within a module or a class. In Kotlin, we can overload functions: there can be multiple declarations of functions that have the same name. Overloaded functions must be distinguishable from each other through their parameter lists. (The types of the parameter list, together with the return type, is known as a function’s signature, but the return type cannot be used to disambiguate overloaded functions.) For example, we can have both of these functions in the same file:

fun square(number: Int) = number * number
fun square(number: Double) = number * number

At the call sites, which function to use is determined from the type of the arguments:

square(4)    // Calls the first function; result is 16 (Int)
square(3.14) // Calls the second function; result is 9.8596 (Double)

While this example happened to use the same expression, that is not necessary - overloaded functions can do completely different things if need be (although your code can get confusing if you make functions that have very different behavior be overloads of each other).

Varargs and optional/named parameters

A function can take an arbitrary number of arguments, similarly to *args in Python, but they must all be of the same type. Unlike Python, you may declare other positional parameters after the variadic one, but there can be at most one variadic parameter. If its type is X, the type of the argument will be XArray if X is a primitive type and Array<X> if not.

fun countAndPrintArgs(vararg numbers: Int) {
    println(numbers.size)
    for (number in numbers) println(number)
}

There are no **kwargs in Kotlin, but you can define optional parameters with default values, and you may choose to name some or all of the parameters when you call the function (whether they’ve got default values or not). Like in Python, the named arguments can be reordered at will:

fun foo(decimal: Double, integer: Int, text: String = "Hello") { ... }

foo(3.14, text = "Bye", integer = 42)
foo(integer = 12, decimal = 3.4)

Unlike in Python, the expressions for the default values are evaluated at function invocation time, not at function definition time, which means that this classic Python trap is safe in Kotlin - every invocation will see a new, empty list:

fun tricky(x: Int, numbers: MutableList<Int> = mutableListOf()) {
    numbers.add(x)
    println(numbers)
}

You can call a variadic function with one array (but not a list or any other iterable) that contains all the variadic arguments, by spreading it with the * operator (same syntax as Python):

val numbers = listOf(1, 2, 3)
countAndPrintArgs(*numbers.toIntArray())

Kotlin has inherited Java’s fidgety array system, so primitive types have got their own array types and conversion functions, while any other type uses the generic Array type, to which you can convert with .toTypedArray().

However, you can’t spread a map into a function call and expect the values in the map to be passed to the parameters named by the keys - the names of the parameters must be known at compile time. If you need runtime-defined parameter names, your function must either take a map or take vararg kwargs: Pair<String, X> (where X is the “lowest common denominator” of the parameter types, in the worst case Any? - be prepared to have to typecast the parameter values, and note that you’ll lose type safety). You can call such a function like this: foo("bar" to 42, "test" to "hello"), since to is an infix function that creates a Pair.

Classes

Kotlin’s object model is substantially different from Python’s. Most importantly, classes are not dynamically modifiable at runtime! (There are some limited exceptions to this, but you generally shouldn’t do it. However, it is possible to dynamically inspect classes and objects at runtime with a feature called reflection - this can be useful, but should be judiciously used.) All properties (attributes) and functions that might ever be needed on a class must be declared either directly in the class body or as extension functions, so you should think carefully through your class design.

Declaration and instantiation

Classes are declared with the class keyword. A basic class without any properties or functions of its own looks like this:

class Empty

You can then create an instance of this class in a way that looks similar to Python, as if the class were a function (but this is just syntactic sugar - unlike Python, classes in Kotlin aren’t really functions):

val object = Empty()

Class names should use UpperCamelCase, just like in Python.

Inherited built-in functions

Every class that doesn’t explicitly declare a parent class inherits from Any, which is the root of the class hierarchy (similar to object in Python) - more on inheritance later. Via Any, every class automatically has the following functions:

Properties

Empty classes aren’t very interesting, so let’s make a class with some properties:

class Person {
    var name = "Anne"
    var age = 32
}

Note that the type of a property must be explicitly specified. As opposed to Python, declaring a property directly inside the class does not create a class-level property, but an instance-level one: every instance of Person will have its own name and age. Their values will start out in every instance as "Anne" and 32, respectively, but the value in each instance can be modified independently of the others:

val a = Person()
val b = Person()
println("${a.age} ${b.age}") // Prints "32 32"
a.age = 42
println("${a.age} ${b.age}") // Prints "42 32"

To be fair, you’d get the same output in Python, but the mechanism would be different: both instances would start out without any attributes of their own (age and name would be attributes on the class), and the first printing would access the class attribute; only the assignment would cause an age attribute to appear on a. In Kotlin, there are no class properties in this example, and each instance starts out with both properties. If you need a class-level property, see the section on companion objects.

Because the set of properties of an object is constrained to be exactly the set of properties that are declared at compile-time in the object’s class, it’s not possible to add new properties to an object or to a class at runtime, so e.g. a.nationality = "Norwegian" won’t compile.

Property names should use lowerCamelCase instead of snake_case.

Constructors and initializer blocks

Properties that don’t have a sensible default should be taken as constructor parameters. Like with Python’s __init__(), Kotlin constructors and initializer blocks run automatically whenever an instance of an object is created (note that there’s nothing that corresponds to __new__()). A Kotlin class may have one primary constructor, whose parameters are supplied after the class name. The primary constructor parameters are available when you initialize properties in the class body, and also in the optional initializer block, which can contain complex initialization logic (a property can be declared without an initial value, in which case it must be initialized in init). Also, you’ll frequently want to use val instead of var in order to make your properties immutable after construction.

class Person(firstName: String, lastName: String, yearOfBirth: Int) {
    val fullName = "$firstName $lastName"
    var age: Int
    
    init {
        age = 2018 - yearOfBirth
    }
}

If all you want to do with a constructor parameter value is to assign it to a property with the same name, you can declare the property in the primary constructor parameter list (the oneliner below is sufficient for both declaring the properties, declaring the constructor parameters, and initializing the properties with the parameters):

class Person(val name: String, var age: Int)

If you need multiple ways to initialize a class, you can create secondary constructors, each of which looks like a function whose name is constructor. Every secondary constructor must invoke another (primary or secondary) constructor by using the this keyword as if it were a function (so that every instance construction eventually calls the primary constructor).

class Person(val name: String, var age: Int) {
    constructor(name: String) : this(name, 0)
    constructor(yearOfBirth: Int, name: String)
        : this(name, 2018 - yearOfBirth)
}

(A secondary constructor can also have a body in curly braces if needs to do more than what the primary constructor does.) The constructors are distinguished from each other through the types of their parameters, like in ordinary function overloading. That’s the reason we had to flip the parameter order in the last secondary constructor - otherwise, it would have been indistinguishable from the primary constructor (parameter names are not a part of a function’s signature and don’t have any effect on overload resolution). In the most recent example, we can now create a Person in three different ways:

val a = Person("Jaime", 35)
val b = Person("Jack") // age = 0
val c = Person(1995, "Lynne") // age = 23

Note that if a class has got a primary constructor, it is no longer possible to create an instance of it without supplying any parameters (unless one of the secondary constructors is parameterless).

Setters and getters

A property is really a backing field (kind of a hidden variable inside the object) and two accessor functions: one that gets the value of the variable and one that sets the value. You can override one or both of the accessors (an accessor that is not overridden automatically gets the default behavior of just returning or setting the backing field directly). Inside an accessor, you can reference the backing field with field. The setter accessor must take a parameter value, which is the value that is being assigned to the property. A getter body could either be a one-line expression preceded by = or a more complex body enclosed in curly braces, while a setter body typically includes an assignment and must therefore be enclosed in curly braces. If you want to validate that the age is nonnegative:

class Person(age: Int) {
    var age = 0
        set(value) {
            if (value < 0) throw IllegalArgumentException(
                    "Age cannot be negative")
            field = value
        }

    init {
        this.age = age
    }
}

Annoyingly, the setter logic is not invoked by the initialization, which instead sets the backing field directly - that’s why we have to use an initializer block in this example in order to verify that newly-created persons also don’t get a negative age. Note the use of this.age in the initializer block in order to distinguish between the identically-named property and constructor parameter.

If for some reason you want to store a different value in the backing field than the value that is being assigned to the property, you’re free to do that, but then you will probably want a getter to give the calling code back what they expect: if you say field = value * 2 in the setter and this.age = age * 2 in the initializer block, you should also have get() = field / 2.

You can also create properties that don’t actually have a backing field, but just reference another property:

val isNewborn
    get() = age == 0

Note that even though this is a read-only property due to declaring it with val (in which case you may not provide a setter), its value can still change since it reads from a mutable property - you just can’t assign to the property. Also, note that the property type is inferred from the return value of the getter.

The indentation in front of the accessors is due to convention; like elsewhere in Kotlin, it has no syntactic significance. The compiler can tell which accessors belong to which properties because the only legal place for an accessor is immediately after the property declaration (and there can be at most one getter and one setter) - so you can’t split the property declaration and the accessor declarations. However, the order of the accessors doesn’t matter.

Member functions

A function declared inside a class is called a member function of that class. Like in Python, every invocation of a member function must be performed on an instance of the class, and the instance will be available during the execution of the function - but unlike Python, the function signature doesn’t declare that: there is no explicit self parameter. Instead, every member function can use the keyword this to reference the current instance, without declaring it. Unlike Python, as long as there is no name conflict with an identically-named parameter or local variable, this can be omitted. If we do this inside a Person class with a name property:

fun present() {
    println("Hello, I'm $name!")
}

We can then do this:

val p = Person("Claire")
p.present() // Prints "Hello, I'm Claire!"

You could have said ${this.name}, but that’s redundant and generally discouraged. Oneliner functions can be declared with an =:

fun greet(other: Person) = println("Hello, ${other.name}, I'm $name!")

Apart from the automatic passing of the instance into this, member functions generally act like ordinary functions.

Because the set of member functions of an object is constrained to be exactly the set of member functions that are declared at compile-time in the object’s class and base classes, it’s not possible to add new member functions to an object or to a class at runtime, so e.g. p.leave = fun() { println("Bye!") } or anything of the sort won’t compile.

Member function names should use lowerCamelCase instead of snake_case.

Lateinit

Kotlin requires that every member property is initialized during instance construction. Sometimes, a class is intended to be used in such a way that the constructor doesn’t have enough information to initialize all properties (such as when making a builder class or when using property-based dependency injection). In order to not have to make those properties nullable, you can use a late-initialized property:

lateinit var name: String

Kotlin will allow you to declare this property without initializing it, and you can set the property value at some point after construction (either directly or via a function). It is the responsibility of the class itself as well as its users to take care not to read the property before it has been set, and Kotlin allows you to write code that reads name as if it were an ordinary, non-nullable property. However, the compiler is unable to enforce correct usage, so if the property is read before it has been set, an UninitializedPropertyAccessException will be thrown at runtime.

Inside the class that declares a lateinit property, you can check if it has been initialized:

if (::name.isInitialized) println(name)

lateinit can only be used with var, not with val, and the type must be non-primitive and non-nullable.

Infix functions

You can designate a one-parameter member function or extension function for use as an infix operator, which can be useful if you’re designing a DSL. The left operand will become this, and the right operand will become the parameter. If you do this inside a Person class that has got a name property:

infix fun marry(spouse: Person) {
    println("$name and ${spouse.name} are getting married!")
}

We can now do this (but it’s still possible to call the function the normal way):

val lisa = Person("Lisa")
val anne = Person("Anne")
lisa marry anne // Prints "Lisa and Anne are getting married!"

All infix functions have the same precedence (which is shared with all the built-in infix functions, such as the bitwise functions and, or, inv, etc.): lower than the arithmetic operators and the .. range operator, but higher than the Elvis operator ?:, comparisons, logic operators, and assignments.

Operators

Most of the operators that are recognized by Kotlin’s syntax have predefined textual names and are available for implementation in your classes, just like you can do with Python’s double-underscore operator names. For example, the binary + operator is called plus. Similarly to the infix example, if you do this inside a Person class that has got a name property:

operator fun plus(spouse: Person) {
    println("$name and ${spouse.name} are getting married!")
}

With lisa and anne from the infix example, you can now do:

lisa + anne // Prints "Lisa and Anne are getting married!"

A particularly interesting operator is the function-call parenthesis pair, whose function name is invoke - if you implement this, you’ll be able to call instances of your class as if they were functions. You can even overload it in order to provide different function signatures.

operator can also be used for certain other predefined functions in order to create fancy effects, such as delegated properties.

Since the available operators are hardcoded into the formal Kotlin syntax, you can not invent new operators, and overriding an operator does not affect its precedence.

Enum classes

Whenever you want a variable that can only take on a limited number of values where the only feature of each value is that it’s distinct from all the other values, you can create an enum class:

enum class ContentKind {
    TOPIC,
    ARTICLE,
    EXERCISE,
    VIDEO,
}

There are exactly four instances of this class, named ContentKind.TOPIC, and so on. Instances of this class can be compared to each other with == and !=, and you can get all the allowable values with ContentKind.values(). You can also tack on more information to each instance if you need:

enum class ContentKind(val kind: String) {
    TOPIC("Topic"),
    ARTICLE("Article"),
    EXERCISE("Exercise"),
    VIDEO("Video"),
    ;

    override fun toString(): String {
        return kind
    }
}

Null safety is enforced as usual, so a variable of type ContentKind can not be null, unlike in Java.

Data classes

Frequently - especially if you want a complex return type from a function or a complex key for a map - you’ll want a quick and dirty class which only contains some properties, but is still comparable for equality and is usable as a map key. If you create a data class, you’ll get automatic implementations of the following functions: toString() (which will produce a string containing all the property names and values), equals() (which will do a per-property equals()), hashCode() (which will hash the individual properties and combine the hashes), and the functions that are required to enable Kotlin to destructure an instance of the class into a declaration (component1(), component2(), etc.):

data class ContentDescriptor(val kind: ContentKind, val id: String) {
    override fun toString(): String {
        return kind.toString() + ":" + id
    }
}

Exceptions

Throwing and catching

Exceptions pretty much work like they do in Python. You throw (raise) one with throw:

throw IllegalArgumentException("Value must be positive")

You catch it with try/catch (which corresponds to try/except in Python):

fun divideOrZero(numerator: Int, denominator: Int): Int {
    try {
        return numerator / denominator
    } catch (e: ArithmeticException) {
        return 0
    }
}

The catch blocks are tried in order until an exception type is found that matches the thrown exception (it doesn’t need to be an exact match; the thrown exception’s class can be a subclass of the declared one), and at most one catch block will be executed. If no match is found, the exception bubbles out of the try/catch.

The finally block (if any) is executed at the end, no matter what the outcome is: either after the try block completes successfully, or after a catch block is executed (even if another exception is thrown by the catch block), or if no matching catch is found.

Unlike Python, try/catch is an expression: the last expression of the try block (if it succeeds) or the chosen catch block becomes the result value (finally doesn’t affect the result), so we can refactor the function body above to:

return try {
    numerator / denominator
} catch (e: ArithmeticException) {
    0
}

The base exception class is Throwable (but it is more common to extend its subclass Exception), and there are a ton of built-in exception classes. If you don’t find one that match your needs, you can create your own by inheriting from an existing exception class.

Note that exceptions are somewhat discouraged in Kotlin except when interacting with Java code. Instead of throwing exceptions in your own code, consider using special return types like Option or Either from the Arrow library.

Nothing

throw is also an expression, and its return type is the special class Nothing, which does not have any instances. The compiler knows that an expression whose type is Nothing will never return normally, and will therefore generally accept its use even where a different type would normally be required, such as after the Elvis operator. If you make a function that always throws, or that starts an infinite loop, you could declare its return type to be Nothing in order to make the compiler aware of this. One fun example of this is the built-in function TODO, which you can call in any expression (possibly supplying a string argument), and it raises a NotImplementedError.

The nullable version Nothing? will be used by the compiler when something is initialized with null and there is no other type information. In val x = null, the type of x will be Nothing?. This type does not have the “never returns normally” semantics; instead, the compiler knows that the value will always be null.

Null safety

Working with nulls

A variable that doesn’t refer to anything refers to null (or you can say that the variable “is null”). As opposed to None in Python, null is not an object - it’s just a keyword that is used to make a variable refer to nothing or to check if it does (that check must be performed with == or !=). Because nulls are a frequent source of programming errors, Kotlin encourages avoiding them as much as possible - a variable cannot actually be null unless it’s been declared to allow for null, which you do by suffixing the type name with ?. For example:

fun test(a: String, b: String?) {
}

The compiler will allow this function to be called as e.g. test("a", "b") or test("a", null), but not as test(null, "b") or test(null, null). Calling test(a, b) is only allowed if the compiler can prove that a cannot possibly be null. Inside of test, the compiler will not allow you to do anything with b that would result in an exception if b should happen to be null - so you can do a.length, but not b.length. However, once you’re inside a conditional where you have checked that b is not null, you can do it:

if (b != null) {
    println(b.length)
}

Or:

if (b == null) {
    // Can't use members of b in here
} else {
    println(b.length)
}

Making frequent null checks is annoying, so if you have to allow for the possibility of nulls, there are several very useful operators in Kotlin to ease working with values that might be null, as described below.

Safe call operator

x?.y evaluates x, and if it is not null, it evaluates x.y (without reevaluating x), whose result becomes the result of the expression - otherwise, you get null. This also works for functions, and it can be chained - for example, x?.y()?.z?.w() will return null if any of x, x.y(), or x.y().z produce null; otherwise, it will return the result of x.y().z.w().

Elvis operator

x ?: y evaluates x, which becomes the result of the expression unless it’s null, in which case you’ll get y instead (which ought to be of a non-nullable type). This is also known as the “Elvis operator”. You can even use it to perform an early return in case of null:

val z = x ?: return y

This will assign x to z if x is non-null, but if it is null, the entire function that contains this expression will stop and return y (this works because return is also an expression, and if it is evaluated, it evaluates its argument and then makes the containing function return the result).

Not-null assertion operator

Sometimes, you’re in a situation where you have a value x that you know is not null, but the compiler doesn’t realize it. This can legitimately happen when you’re interacting with Java code, but if it happens because your code’s logic is more complicated than the compiler’s ability to reason about it, you should probably restructure your code. If you can’t convince the compiler, you can resort to saying x!! to form an expression that produces the value of x, but whose type is non-nullable:

val x: String? = javaFunctionThatYouKnowReturnsNonNull()
val y: String = x!!

It can of course be done as a single expression: val x = javaFunctionThatYouKnowReturnsNonNull()!!.

!! will will raise a NullPointerException if the value actually is null. So you could also use it if you really need to call a particular function and would rather have an exception if there’s no object to call it on (maybeNull!!.importantFunction()), although a better solution (because an NPE isn’t very informational) is this:

val y: String = x ?: throw SpecificException("Useful message")
y.importantFunction()

The above could also be a oneliner - and note that the compiler knows that because the throw will prevent y from coming into existence if x is null, y must be non-null if we reach the line below. Contrast this with x?.importantFunction(), which is a no-op if x is null.

Functional programming

Function types

Like in Python, functions in Kotlin are first-class values - they can be assigned to variables and passed around as parameters. The type a function is a function type, which is indicated with a parenthesized parameter type list and an arrow to the return type. Consider this function:

fun safeDivide(numerator: Int, denominator: Int) =
    if (denominator == 0.0) 0.0 else numerator.toDouble() / denominator

It takes two Int parameters and returns a Double, so its type is (Int, Int) -> Double. We can reference the function itself by prefixing its name with ::, and we can assign it to a variable (whose type would normally be inferred, but we show the type signature for demonstration):

val f: (Int, Int) -> Double = ::safeDivide

When you have a variable or parameter of function type (sometimes called a function reference), you can call it as if it were an ordinary function, and that will cause the referenced function to be called:

val quotient = f(3.14, 0.0)

It is possible for a class to implement a function type as if it were an interface. It must then supply an operator function called invoke with the given signature, and instances of that class may then be assigned to a variable of that function type:

class Divider : (Int, Int) -> Double {
    override fun invoke(numerator: Int, denominator: Int): Double = ...
}

Function literals: lambda expressions and anonymous functions

Like in Python, you can write lambda expressions: unnamed function declarations with a very compact syntax, which evaluate to callable function objects. In Kotlin, lambdas can contain multiple statements, which make them useful for more complex tasks than the single-expression lambdas of Python. The last statement must be an expression, whose result will become the return value of the lambda (unless Unit is the return type of the variable/parameter that the lambda expression is assigned to, in which case the lambda has no return value). A lambda expression is enclosed in curly braces, and begins by listing its parameter names and possibly their types (unless the types can be inferred from context):

val safeDivide = { numerator: Int, denominator: Int ->
    if (denominator == 0.0) 0.0 else numerator.toDouble() / denominator
}

The type of safeDivide is (Int, Int) -> Double. Note that unlike function type declarations, the parameter list of a lambda expression must not be enclosed in parentheses.

Note that the other uses of curly braces in Kotlin, such as in function and class definitions and after if/else/for/while statements, are not lambda expressions (so it is not the case that if is a function that conditionally executes a lambda function).

The return type of a lambda expression is inferred from the type of the last expression inside it (or from the function type of the variable/parameter that the lambda expression is assigned to). If a lambda expression is passed as a function parameter (which is the ordinary use) or assigned to a variable with a declared type, Kotlin can infer the parameter types too, and you only need to specify their names:

val safeDivide: (Int, Int) -> Double = { numerator, denominator ->
    if (denominator == 0) 0.0 else numerator.toDouble() / denominator
}

Or:

fun callAndPrint(function: (Int, Int) -> Double) {
    println(function(2, 0))
}

callAndPrint({ numerator, denominator ->
    if (denominator == 0) 0.0 else numerator.toDouble() / denominator
})

A parameterless lambda does not need the arrow. A one-parameter lambda can choose to omit the parameter name and the arrow, in which case the parameter will be called it:

val square: (Double) -> Double = { it * it }

If the type of the last parameter to a function is a function type and you want to supply a lambda expression, you can place the lambda expression outside of the parameter parentheses. If the lambda expression is the only parameter, you can omit the parentheses entirely. This is very useful for constructing DSLs.

fun callWithPi(function: (Double) -> Double) {
    println(function(3.14))
}

callWithPi { it * it }

If you want to be more explicit about the fact that you’re creating a function, you can make an anonymous function, which is still an expression rather than a declaration:

callWithPi(fun(x: Double): Double { return x * x })

Or:

callWithPi(fun(x: Double) = x * x)

Lambda expressions and anonymous functions are collectively called function literals.

Comprehensions

Kotlin can get quite close to the compactness of Python’s list/dict/set comprehensions. Assuming that people is a collection of Person objects with a name property:

val shortGreetings = people
    .filter { it.name.length < 10 }
    .map { "Hello, ${it.name}!" }

corresponds to

short_greetings = [
    "Hello, %s!" % p.name
    for p in people
    if len(p.name) < 10
]

In some ways, this is easier to read because the operations are specified in the order they are applied to the values. The result will be an immutable List<T>, where T is whichever type is produced by the transformations you use (in this case, String). If you need a mutable list, call toMutableList() at the end. If you want a set, call toSet() or toMutableSet() at the end. If you want to transform a collection into a map, call associateBy(), which takes two lambdas that specify how to extract the key and the value from each element: people.associateBy({it.ssn}, {it.name}) (you can omit the second lambda if you want the entire element to be the value, and you can call toMutableMap() at the end if you want the result to be mutable).

These transformations can also be applied to Sequence<T>, which is similar to Python’s generators and allows for lazy evaluation. If you have a huge list and you want to process it lazily, you can call asSequence() on it.

There’s a vast collection of functional programming-style operations available in the kotlin.collections package.

Receivers

The signature of a member function or an extension function begins with a receiver: the type upon which the function can be invoked. For example, the signature of toString() is Any.() -> String - it can be called on any non-null object (the receiver), it takes no parameters, and it returns a String. It is possible to write a lambda function with such a signature - this is called a function literal with receiver, and is extremely useful for building DSLs.

A function literal with receiver is perhaps easiest to think of as an extension function in the form of a lambda expression. The declaration looks like an ordinary lambda expression; what makes it take a receiver is the context - it must be passed to a function that takes a function with receiver as a parameter, or assigned to a variable/property whose type is a function type with receiver. The only way to use a function with receiver is to invoke it on an instance of the receiver class, as if it were a member function or extension function. For example:

class Car(val horsepowers: Int)

val boast: Car.() -> String = { "I'm a car with $horsepowers HP!"}

val car = Car(120)
println(car.boast())

Inside a lambda expression with receiver, you can use this to refer to the receiver object (in this case, car). As usual, you can omit this if there are no naming conflicts, which is why we can simply say $horsepowers instead of ${this.horsepowers}. So beware that in Kotlin, this can have different meanings depending on the context: if used inside (possibly nested) lambda expressions with receivers, it refers to the receiver object of the innermost enclosing lambda expression with receiver. If you need to “break out” of the function literal and get the “original” this (the instance the member function you’re inside is executing on), mention the containing class name after this@ - so if you’re inside a function literal with receiver inside a member function of Car, use this@Car.

As with other function literals, if the function takes one parameter (other than the receiver object that it is invoked on), the single parameter is implicitly called it, unless you declare another name. If it takes more than one parameter, you must declare their names.

Here’s a small DSL example for constructing tree structures:

class TreeNode(val name: String) {
    val children = mutableListOf<TreeNode>()

    fun node(name: String, initialize: (TreeNode.() -> Unit)? = null) {
        val child = TreeNode(name)
        children.add(child)
        if (initialize != null) {
            child.initialize()
        }
    }
}

fun tree(name: String, initialize: (TreeNode.() -> Unit)? = null): TreeNode {
    val root = TreeNode(name)
    if (initialize != null) {
        root.initialize()
    }
    return root
}

val t = tree("root") {
    node("math") {
        node("algebra")
        node("trigonometry")
    }
    node("science") {
        node("physics")
    }
}

The block after tree("root") is the first function literal with receiver, which will be passed to tree() as the initialize parameter. According to the parameter list of tree(), the receiver is of type TreeNode, and therefore, tree() can call initialize() on root. root then becomes this inside the scope of that lambda expression, so when we call node("math"), it implicitly says this.node("math"), where this refers to the same TreeNode as root. The next block is passed to TreeNode.node(), and is invoked on the first child of the root node, namely math, and inside it, this will refer to math.

If we had wanted to express the same thing in Python, it would have looked like this, and we would be hamstrung by the fact that lambda functions can only contain one expression, so we need explicit function definitions for everything but the oneliners:

class TreeNode:
    def __init__(self, name):
        self.name = name
        self.children = []

    def node(self, name, initialize=None):
        child = TreeNode(name)
        self.children.append(child)
        if initialize:
            initialize(child)

def tree(name, initialize=None):
    root = TreeNode(name)
    if initialize:
        initialize(root)
    return root

def init_root(root):
    root.node("math", init_math)
    root.node("science",
              lambda science: science.node("physics"))

def init_math(math):
    math.node("algebra")
    math.node("trigonometry")

t = tree("root", init_root)

The official docs also have a very cool example with a DSL for constructing HTML documents.

Inline functions

There’s a little bit of runtime overhead associated with lambda functions: they are really objects, so they must be instantiated, and (like other functions) calling them takes a little bit of time too. If we use the inline keyword on a function, we tell the compiler to inline both the function and its lambda parameters (if any) - that is, the compiler will copy the code of the function (and its lambda parameters) into every callsite, thus eliminating the overhead of both the lambda instantiation and the calling of the function and the lambdas. This will happen unconditionally, unlike in C and C++, where inline is more of a hint to the compiler. This will cause the size of the compiled code to grow, but it may be worth it for certain small but frequently-called functions.

inline fun time(action: () -> Unit): Long {
    val start = Instant.now().toEpochMilli()
    action()
    return Instant.now().toEpochMilli() - start
}

Now, if you do:

val t = time { println("Lots of code") }
println(t)

The compiler will generate something like this (except that start won’t collide with any other identifiers with the same name):

val start = Instant.now().toEpochMilli()
println("Lots of code")
val t = Instant.now().toEpochMilli() - start
println(t)

In an inline function definition, you can use noinline in front of any function-typed parameter to prevent the lambda that will be passed to it from also being inlined.

Nice utility functions

run(), let(), and with()

?. is nice if you want to call a function on something that might be null. But what if you want to call a function that takes a non-null parameter, but the value you want to pass for that parameter might be null? Try run(), which is an extension function on Any? that takes a lambda with receiver as a parameter and invokes it on the value that it’s called on, and use ?. to call run() only if the object is non-null:

val result = maybeNull?.run { functionThatCanNotHandleNull(this) }

If maybeNull is null, the function won’t be called, and result will be null; otherwise, it will be the return value of functionThatCanNotHandleNull(this), where this refers to maybeNull. You can chain run() calls with ?. - each one will be called on the previous result if it’s not null:

val result = maybeNull
    ?.run { firstFunction(this) }
    ?.run { secondFunction(this) }

The first this refers to maybeNull, the second one refers to the result of firstFunction(), and result will be the result of secondFunction() (or null if maybeNull or any of the intermediate results were null).

A syntactic variation of run() is let(), which takes an ordinary function type instead of a function type with receiver, so the expression that might be null will be referred to as it instead of this.

Both run() and let() are also useful if you’ve got an expression that you need to use multiple times, but you don’t care to come up with a variable name for it and make a null check:

val result = someExpression?.let {
   firstFunction(it)
   it.memberFunction() + it.memberProperty
}

Yet another version is with(), which you can also use to avoid coming up with a variable name for an expression, but only if you know that its result will be non-null:

val result = with(someExpression) {
   firstFunction(this)
   memberFunction() + memberProperty
}

In the last line, there’s an implicit this. in front of both memberFunction() and memberProperty (if these exist on the type of someExpression). The return value is that of the last expression.

apply() and also()

If you don’t care about the return value from the function, but you want to make one or more calls involving something that might be null and then keep on using that value, try apply(), which returns the value it’s called on. This is particularly useful if you want to work with many members of the object in question:

maybeNull?.apply {
    firstFunction(this)
    secondFunction(this)
    memberPropertyA = memberPropertyB + memberFunctionA()
}?.memberFunctionB()

Inside the apply block, this refers to maybeNull. There’s an implicit this in front of memberPropertyA, memberPropertyB, and memberFunctionA (unless these don’t exist on maybeNull, in which case they’ll be looked for in the containing scopes). Afterwards, memberFunctionB() is also invoked on maybeNull.

If you find the this syntax to be confusing, you can use also instead, which takes ordinary lambdas:

maybeNull?.also {
    firstFunction(it)
    secondFunction(it)
    it.memberPropertyA = it.memberPropertyB + it.memberFunctionA()
}?.memberFunctionB()

takeIf() and takeUnless()

If you want to use a value only if it satisfies a certain condition, try takeIf(), which returns the value it’s called on if it satisfies the given predicate, and null otherwise. There’s also takeUnless(), which inverts the logic. You can follow this with a ?. to perform an operation on the value only if it satisfies the predicate. Below, we compute the square of some expression, but only if the expression value is at least 42:

val result = someExpression.takeIf { it >= 42 } ?.let { it * it }

Packages and imports

Packages

Every Kotlin file should belong to a package. This is somewhat similar to modules in Python, but files need to explicitly declare which package they belong to, and a package implicitly comes into existence whenever any file declares itself to belong to that package (as opposed to explicitly defining a module with __init__.py and having all the files in that directory implicitly belong to the module). The package declaration must go on the top of the file:

package content.exercises

If a file doesn’t declare a package, it belongs to the nameless default package. This should be avoided, as it will make it hard to reference the symbols from that file in case of naming conflicts (you can’t explicitly import the empty package).

Package names customarily correspond to the directory structure - note that the source file name should not be a part of the package name (so if you follow this, file-level symbol names must be unique within an entire directory, not just within a file). However, this correspondence is not required, so if you’re going to do interop with Java code and all your package names must start with the same prefix, e.g. org.khanacademy, you might be relieved to learn that you don’t need to put all your code inside org/khanacademy (which is what Java would have forced you to do) - instead, you could start out with a directory called e.g. content, and the files inside it could declare that they belong to the package org.khanacademy.content. However, if you have a mixed project with both Kotlin and Java code, the convention is to use the Java-style package directories for Kotlin code too.

While the dots suggest that packages are nested inside each other, that’s not actually the case from a language standpoint. While it’s a good idea to organize your code such that the “subpackages” of content, such as content.exercises and content.articles, both contain content-related code, these three packages are unrelated from a language standpoint. However, if you use modules (as defined by your build system), it is typically the case that all “subpackages” go in the same module, in which case symbols with internal visibility are visible throughout the subpackages.

Package names customarily contain only lowercase letters (no underscores) and the separating dots.

Imports

In order to use something from a package, it is sufficient to use the package name to fully qualify the name of the symbol at the place where you use the symbol:

val exercise = content.exercises.Exercise()

This quickly gets unwieldy, so you will typically import the symbols you need. You can import a specific symbol:

import content.exercises.Exercise

Or an entire package, which will bring in all the symbols from that package:

import content.exercises.*

With either version of the import, you can now simply do:

val exercise = Exercise()

If there is a naming conflict, you should usually import just one of the symbols and fully qualify the usages of the other. If both are heavily used, you can rename the symbol at import time:

import content.exercises.Exercise as Ex

In Kotlin, importing is a compile-time concept - importing something does not actually cause any code to run (unlike Python, where all top-level statements in a file are executed at import time). Therefore, circular imports are allowed, but they might suggest a design problem in your code. However, during execution, a class will be loaded the first time it (or any of its properties or functions) is referenced, and class loading causes companion objects to be initialized - this can lead to runtime exceptions if you have circular dependencies.

Every file implicitly imports its own package and a number of built-in Kotlin and Java packages.

Visibility modifiers

Kotlin allows you to enforce symbol visibility (which Python only does via underscore conventions) via visibility modifiers, which can be placed on symbol declarations. If you don’t supply a visibility modifier, you get the default visibility level, which is public.

The meaning of a visibility modifier depends on whether it’s applied to a top-level declaration or to a declaration inside a class. For top-level declarations:

For example, private could be used to define a property or helper function that is needed by several functions in one file, or a complex type returned by one of your private functions, without leaking them to the rest of the codebase:

private class ReturnType(val a: Int, val b: Double, val c: String)

private fun secretHelper(x: Int) = x * x

private const val secretValue = 3.14

For a symbol that is declared inside a class:

A constructor can also have a visibility modifier. If you want to place one on the primary constructor (which you might want to do if you have a number of secondary constructors which all invoke a complicated primary constructor that you don’t want to expose), you need to include the constructor keyword: class Person private constructor(val name: String).

Visibility modifiers can’t be placed on local variables, since their visibility is always limited to the containing block.

The type of a property, and the types that are used for the parameters and the return type of a function, must be “at least as visible” as the property/function itself. For example, a public function can’t take a private type as a parameter.

The visibility level only affects the lexical visibility of the symbol - i.e., where the compiler allows you to type out the symbol. It does not affect where instances are used: for example, a public top-level function may well return an instance of a private class, as long as the return type doesn’t mention the private class name but is instead a public base class of the private class (possibly Any) or a public interface that the private class implements. When you subclass a class, its private members are also inherited by the subclass, but are not directly accessible there - however, if you call an inherited public function that happens to access a private member, that’s fine.

Inheritance

Subclassing

Kotlin supports single-parent class inheritance - so each class (except the root class Any) has got exactly one parent class, called a superclass. Kotlin wants you to think through your class design to make sure that it’s actually safe to subclass it, so classes are closed by default and can’t be inherited from unless you explicitly declare the class to be open or abstract. You can then subclass from that class by declaring a new class which mentions its parent class after a colon:

open class MotorVehicle
class Car : MotorVehicle()

Classes that don’t declare a superclass implicitly inherit from Any. The subclass must invoke one of the constructors of the base class, passing either parameters from its own constructor or constant values:

open class MotorVehicle(val maxSpeed: Double, val horsepowers: Int)
class Car(
    val seatCount: Int,
    maxSpeed: Double
) : MotorVehicle(maxSpeed, 100)

The subclass inherits all members that exist in its superclass - both those that are directly defined in the superclass and the ones that the superclass itself has inherited. In this example, Car contains the following members:

Note that the terms “subclass” and “superclass” can span multiple levels of inheritance - Car is a subclass of Any, and Any is the superclass of everything. If we want to restrict ourselves to one level of inheritance, we will say “direct subclass” or “direct superclass”.

Note that we do not use val in front of maxSpeed in Car - doing so would have introduced a distinct property in Car that would have shadowed the one inherited from MotorVehicle. As written, it’s just a constructor parameter that we pass on to the superconstructor.

private members (and internal members from superclasses in other modules) are also inherited, but are not directly accessible: if the superclass contains a private property foo that is referenced by a public function bar(), instances of the subclass will contain a foo; they can’t use it directly, but they are allowed to call bar().

When an instance of a subclass is constructed, the superclass “part” is constructed first (via the superclass constructor). This means that during execution of the constructor of an open class, it could be that the object being constructed is an instance of a subclass, in which case the subclass-specific properties have not been initialized yet. For that reason, calling an open function from a constructor is risky: it might be overridden in the subclass, and if it is accessing subclass-specific properties, those won’t be initialized yet.

Overriding

If a member function or property is declared as open, subclasses may override it by providing a new implementation. Let’s say that MotorVehicle declares this function:

open fun drive() =
    "$horsepowers HP motor vehicle driving at $maxSpeed MPH"

If Car does nothing, it will inherit this function as-is, and it will return a message with the car’s horsepowers and max speed. If we want a car-specific message, Car can override the function by redeclaring it with the override keyword:

override fun drive() =
   "$seatCount-seat car driving at $maxSpeed MPH"

The signature of the overriding function must exactly match the overridden one, except that the return type in the overriding function may be a subtype of the return type of the overridden function.

If what the overriding function wants to do is an extension of what the overridden function did, you can call the overridden function via super (either before, after, or between other code):

override fun drive() =
    super.drive() + " with $seatCount seats"

Interfaces

The single-parent rule often becomes too limiting, as you’ll often find commonalities between classes in different branches of a class hierarchy. These commonalities can be expressed in interfaces.

An interface is essentially a contract that a class may choose to sign; if it does, the class is obliged to provide implementations of the properties and functions of the interface. However, an interface may (but typically doesn’t) provide a default implementation of some or all of its properties and functions. If a property or function has a default implementation, the class may choose to override it, but it doesn’t have to. Here’s an interface without any default implementations:

interface Driveable {
    val maxSpeed: Double
    fun drive(): String
}

We can choose to let MotorVehicle implement that interface, since it’s got the required members - but now we need to mark those members with override, and we can remove open since an overridden function is implicitly open:

open class MotorVehicle(
    override val maxSpeed: Double,
    val wheelCount: Int
) : Driveable {
    override fun drive() = "Wroom!"
}

If we were to introduce another class Bicycle, which should be neither a subclass nor a superclass of MotorVehicle, we could still make it implement Driveable, as long as we declare maxSpeed and drive in Bicycle.

Subclasses of a class that implements an interface (in this case, Car) are also considered to be implementing the interface.

A symbol that is declared inside an interface normally should be public. The only other legal visibility modifier is private, which can only be used if the function body is supplied - that function may then be called by each class that implements the interface, but not by anyone else.

As for why you would want to create an interface, other than as a reminder to have your classes implement certain members, see the section on polymorphism.

Abstract classes

Some superclasses are very useful as a grouping mechanism for related classes and for providing shared functions, but are so general that they’re not useful on their own. MotorVehicle seems to fit this description. Such a class should be declared abstract, which will prevent the class from being instantiated directly:

abstract class MotorVehicle(val maxSpeed: Double, val wheelCount: Int)

Now, you can no longer say val mv = MotorVehicle(100, 4).

Abstract classes are implicitly open, since they are useless if they don’t have any concrete subclasses.

When an abstract class implements one or more interfaces, it is not required to provide definitions of the members of its interfaces (but it can if it wants to). It must still declare such members, using abstract override and not providing any body for the function or property:

abstract override val foo: String
abstract override fun bar(): Int

Being abstract is the only way to “escape” from having to implement the members of your interfaces, by offloading the work onto your subclasses - if a subclass wants to be concrete, it must implement all the “missing” members.

Polymorphism

Polymorphism is the ability to treat objects with similar traits in a common way. In Python, this is achieved via ducktyping: if x refers to some object, you can call x.quack() as long as the object happens to have the function quack() - nothing else needs to be known (or rather, assumed) about the object. That’s very flexible, but also risky: if x is a parameter, every caller of your function must be aware that the object they pass to it must have quack(), and if someone gets it wrong, the program blows up at runtime.

In Kotlin, polymorphism is achieved via the class hierarchy, in such a way that it is impossible to run into a situation where a property or function is missing. The basic rule is that a variable/property/parameter whose declared type is A may refer to an instance of a class B if and only if B is a subtype of A. This means that either, A must be a class and B must be a subclass of A, or that A must be an interface and B must be a class that implements that interface or be a subclass of a class that does. With our classes and interfaces from the previous sections, we can define these functions:

fun boast(mv: MotorVehicle) =
    "My ${mv.wheelCount} wheel vehicle can drive at ${mv.maxSpeed} MPH!"

fun ride(d: Driveable) =
    "I'm riding my ${d.drive()}"

and call them like this:

val car = Car(4, 120)
boast(car)
ride(car)

We’re allowed to pass a Car to boast() because Car is a subclass of MotorVehicle. We’re allowed to pass a Car to ride() because Car implements Driveable (thanks to being a subclass MotorVehicle). Inside boast(), we’re only allowed to access the members of the declared parameter type MotorVehicle, even if we’re in a situation where we know that it’s really a Car (because there could be other callers that pass a non-Car). Inside ride(), we’re only allowed to access the members of the declared parameter type Driveable. This ensures that every member lookup is safe - the compiler only allows you to pass objects that are guaranteed to have the necessary members. The downside is that you will sometimes be forced to declare “unnecessary” interfaces or wrapper classes in order to make a function accept instances of different classes.

With collections and functions, polymorphism becomes more complicated - see the section on generics.

Casting and type testing

When you take an interface or an open class as a parameter, you generally don’t know the real type of the parameter at runtime, since it could be an instance of a subclass or of any class that implements the interface. It is possible to check what the exact type is, but like in Python, you should generally avoid it and instead design your class hierarchy such that you can do what you need by proper overriding of functions or properties.

If there’s no nice way around it, and you need to take special actions based on what type something is or to access functions/properties that only exist on some classes, you can use is to check if the real type of an object is a particular class or a subclass thereof (or an implementor of an interface). When this is used as the condition in an if, the compiler will let you perform type-specific operations on the object inside the if body:

fun foo(x: Any) {
    if (x is Person) {
        println("${x.name}") // This wouldn't compile outside the if
    }
}

If you want to check for not being an instance of a type, use !is. Note that null is never an instance of any non-nullable type, but it is always an “instance” of any nullable type (even though it technically isn’t an instance, but an absence of any instance).

The compiler will not let you perform checks that can’t possibly succeed because the declared type of the variable is a class that is on an unrelated branch of the class hierarchy from the class you’re checking against - if the declared type of x is MotorVehicle, you can’t check if x is a Person. If the right-hand side of is is an interface, Kotlin will allow the type of the left-hand side to be any interface or open class, because it could be that some subclass thereof implements the interface.

If your code is too clever for the compiler, and you know without the help of is that x is an instance of Person but the compiler doesn’t, you can cast your value with as:

val p = x as Person

This will raise a ClassCastException if the object is not actually an instance of Person or any of its subclasses. If you’re not sure what x is, but you’re happy to get null if it’s not a Person, you can use as?, which will return null if the cast fails. Note that the resulting type is Person?:

val p = x as? Person

You can also use as to cast to a nullable type. The difference between this and the previous as? cast is that this one will fail if x is a non-null instance of another type than Person:

val p = x as Person?

Delegation

If you find that an interface that you want a class to implement is already implemented by one of the properties of the class, you can delegate the implementation of that interface to that property with by:

interface PowerSource {
    val horsepowers: Int
}

class Engine(override val horsepowers: Int) : PowerSource

open class MotorVehicle(val engine: Engine): PowerSource by engine

This will automatically implement all the interface members of PowerSource in MotorVehicle by invoking the same member on engine. This only works for properties that are declared in the constructor.

Delegated properties

Let’s say that you’re writing a simple ORM. Your database library represents a row as instances of a class Entity, with functions like getString("name") and getLong("age") for getting typed values from the given columns. We could create a typed wrapper class like this:

abstract class DbModel(val entity: Entity)

class Person(val entity: Entity) : DbModel(entity) {
    val name = entity.getString("name")
    val age = entity.getLong("age")
}

That was easy, but maybe we’d want to do lazy-loading so that we won’t spend time on extracting the fields that won’t be used (especially if some of them contain a lot of data in a format that it is time-consuming to parse), and maybe we’d like support for default values. While we could implement that logic in a get() block, it would need to be duplicated in every property. Alternatively, we could implement the logic in a separate StringProperty class (note that this simple example is not thread-safe):

class StringProperty(
    private val model: DbModel,
    private val fieldName: String,
    private val defaultValue: String? = null
) {
    private var _value: String? = defaultValue
    private var loaded = false
    val value: String?
        get() {
            // Warning: This is not thread-safe!
            if (loaded) return _value
            if (model.entity.contains(fieldName)) {
                _value = model.entity.getString(fieldName)
            }
            loaded = true
            return _value
        }
}

// In Person
val name = StringProperty(this, "name", "Unknown Name")

Unfortunately, using this would require us to type p.name.value every time we wanted to use the property. We could do the following, but that’s also not great since it introduces an extra property:

// In Person
private val _name = StringProperty(this, "name", "Unknown Name")
val name get() = _name.value

The solution is a delegated property, which allows you to specify the behavior of getting and setting a property (somewhat similar to implementing __getattribute__() and __setattribute__() in Python, but for one property at a time).

class DelegatedStringProperty(
    private val fieldName: String,
    private val defaultValue: String? = null
) {
    private var _value: String? = null
    private var loaded = false
    operator fun getValue(thisRef: DbModel, property: KProperty<*>): String? {
        if (loaded) return _value
        if (thisRef.entity.contains(fieldName)) {
            _value = thisRef.entity.getString(fieldName)
        }
        loaded = true
        return _value
    }
}

The delegated property can be used like this to declare a property in Person - note the use of by instead of =:

val name by DelegatedStringProperty(this, "name", "Unknown Name")

Now, whenever anyone reads p.name, getValue() will be invoked with p as thisRef and metadata about the name property as property. Since thisRef is a DbModel, this delegated property can only be used inside DbModel and its subclasses.

A nice built-in delegated property is lazy, which is a properly threadsafe implementation of the lazy loading pattern. The supplied lambda expression will only be evaluated once, the first time the property is accessed.

val name: String? by lazy {
    if (thisRef.entity.contains(fieldName)) {
        thisRef.entity.getString(fieldName)
    } else null
}

Sealed classes

If you want to restrict the set of subclasses of a base class, you can declare the base class to be sealed (which also makes it abstract), in which case you can only declare subclasses in the same file. The compiler then knows the complete set of possible subclasses, which will let you do exhaustive when expression for all the possible subtypes without the need for an else clause (and if you add another subclass in the future and forget to update the when, the compiler will let you know).

Objects and companion objects

Object declarations

If you need a singleton - a class that only has got one instance - you can declare the class in the usual way, but use the object keyword instead of class:

object CarFactory {
    val cars = mutableListOf<Car>()
    
    fun makeCar(horsepowers: Int): Car {
        val car = Car(horsepowers)
        cars.add(car)
        return car
    }
}

There will only ever be one instance of this class, and the instance (which is created the first time it is accessed, in a thread-safe manner) has got the same name as the class:

val car = CarFactory.makeCar(150)
println(CarFactory.cars.size)

Companion objects

If you need a function or a property to be tied to a class rather than to instances of it (similar to @staticmethod in Python), you can declare it inside a companion object:

class Car(val horsepowers: Int) {
    companion object Factory {
        val cars = mutableListOf<Car>()

        fun makeCar(horsepowers: Int): Car {
            val car = Car(horsepowers)
            cars.add(car)
            return car
        }
    }
}

The companion object is a singleton, and its members can be accessed directly via the name of the containing class (although you can also insert the name of the companion object if you want to be explicit about accessing the companion object):

val car = Car.makeCar(150)
println(Car.Factory.cars.size)

In spite of this syntactical convenience, the companion object is a proper object on its own, and can have its own supertypes - and you can assign it to a variable and pass it around. If you’re integrating with Java code and need a true static member, you can annotate a member inside a companion object with @JvmStatic.

A companion object is initialized when the class is loaded (typically the first time it’s referenced by other code that is being executed), in a thread-safe manner. You can omit the name, in which case the name defaults to Companion. A class can only have one companion object, and companion objects can not be nested.

Companion objects and their members can only be accessed via the containing class name, not via instances of the containing class. Kotlin does not support class-level functions that also can be overridden in subclasses (like @classmethod in Python). If you try to redeclare a companion object in a subclass, you’ll just shadow the one from the base class. If you need an overridable “class-level” function, make it an ordinary open function in which you do not access any instance members - you can override it in subclasses, and when you call it via an object instance, the override in the object’s class will be called. It is possible, but inconvenient, to call functions via a class reference in Kotlin, so we won’t cover that here.

Object expressions

Java only got support for function types and lambda expressions a few years ago. Previously, Java worked around this by using an interface to define a function signature and allowing an inline, anonymous definition of a class that implements the interface. This is also available in Kotlin, partly for compatibility with Java libraries and partly because it can be handy for specifying event handlers (in particular if there is more than one event type that must be listened for by the same listener object). Consider an interface or a (possibly abstract) class, as well a function that takes an instance of it:

interface Vehicle {
    fun drive(): String
}

fun start(vehicle: Vehicle) = println(vehicle.drive())

By using an object expression, you can now define an anonymous, unnamed class and at the same time create one instance of it, called an anonymous object:

start(object : Vehicle {
    override fun drive() = "Driving really fast"
})

If the supertype has a constructor, it must be invoked with parentheses after the supertype name. You can specify multiple supertypes if need be (but as usual, at most one superclass).

Since an anonymous class has no name, it can’t be used as a return type - if you do return an anonymous object, the function’s return type must be Any.

In spite of the object keyword being used, a new instance of the anonymous class will be created whenever the object expression is evaluated.

The body of an object expression may access, and possibly modify, the local variables of the containing scope.

Generics

Generic type parameters

One might think that static typing would make it very impractical to make collection classes or any other class that needs to contain members whose types vary with each usage. Generics to the rescue: they allow you to specify a “placeholder” type in a class or function that must be filled in whenever the class or function is used. For example, a node in a linked list needs to contain data of some type that is not known when we write the class, so we introduce a generic type parameter T (they are conventionally given single-letter names):

class TreeNode<T>(val value: T?, val next: TreeNode<T>? = null)

Whenever you create an instance of this class, you must specify an actual type in place of T, unless the compiler can infer it from the constructor parameters: TreeNode("foo") or TreeNode<String>(null). Every use of this instance will act as if it were an instance of a class that looks like this:

class TreeNode<String>(val value: String?, val next: TreeNode<String>? = null)

Member properties and member functions inside a generic class may for the most part use the class’ generic type parameters as if they were ordinary types, without having to redeclare them. It is also possible to make functions that take more generic parameters than the class does, and to make generic functions inside nongeneric classes, and to make generic top-level functions (which is what we’ll do in the next example). Note the different placement of the generic type parameter in generic function declarations:

fun <T> makeLinkedList(vararg elements: T): TreeNode<T>? {
    var node: TreeNode<T>? = null
    for (element in elements.reversed()) {
        node = TreeNode(element, node)
    }
    return node
}

Constraints

You can restrict the types that can be used for a generic type parameter, by specifying that it must be an instance of a specific type or of a subclass thereof. If you’ve got a class or interface called Vehicle, you can do:

class TreeNode<T : Vehicle>

Now, you may not create a TreeNode of a type that is not a subclass/implementor of Vehicle. Inside the class, whenever you’ve got a value of type T, you may access all the public members of Vehicle on it.

If you want to impose additional constraints, you must use a separate where clause, in which case the type parameter must be a subclass of the given class (if you specify a class, and you can specify at most one) and implement all the given interfaces. You may then access all the public members of all the given types whenever you’ve got a value of type T:

class TreeNode<T> where T : Vehicle, T : HasWheels

Variance

Introduction

Pop quiz: if Apple is a subtype of Fruit, and Bowl is a generic container class, is Bowl<Apple> a subtype of Bowl<Fruit>? The answer is - perhaps surprisingly - no. The reason is that if it were a subtype, we would be able to break the type system like this:

fun add(bowl: Bowl<Fruit>, fruit: Fruit) = bowl.add(fruit)

val bowl = Bowl<Apple>()
add(bowl, Pear()) // Doesn't actually compile!
val apple = bowl.get() // Boom!

If the second-to-last line compiled, it would allow us to put a pear into what is ostensibly a bowl of only apples, and your code would explode when it tried to extract the “apple” from the bowl. However, it’s frequently useful to be able to let the type hierarchy of a generic type parameter “flow” to the generic class. As we saw above, though, some care must be taken - the solution is to restrict the direction in which you can move data in and out of the generic object.

Declaration-site covariance and contravariance

If you have an instance of Generic<Subtype>, and you want to refer to it as a Generic<Supertype>, you can safely get instances of the generic type parameter from it - these will truly be instances of Subtype (because they come from an instance of Generic<Subtype>), but they will appear to you as instances of Supertype (because you’ve told the compiler that you have a Generic<Supertype>). This is safe; it is called covariance, and Kotlin lets you do declaration-site covariance by putting out in front of the generic type parameter. If you do, you may only use that type parameter as a return type, not as a parameter type. Here is the simplest useful covariant interface:

interface Producer<out T> {
    fun get(): T
}

It is safe to treat a Producer<Apple> as if it were a Producer<Fruit> - the only thing it will ever produce is Apple instances, but that’s okay, because an Apple is a Fruit.

Conversely, if you have an instance of Generic<Supertype>, and you want to refer to it as a Generic<Subtype> (which you can’t do with nongeneric classes), you can safely give instances of the generic type parameter to it - the compiler will require those instances to be of the type Subtype, which will be acceptable to the real instance because it can handle any Supertype. This is called contravariance, and Kotlin lets you do declaration-site contravariance by putting in in front of the generic type parameter. If you do, you may only use that type parameter as a parameter type, not as a return type. Here is the simplest useful contravariant interface:

interface Consumer<in T> {
    fun add(item: T)
}

It is safe to treat a Consumer<Fruit> as a Consumer<Apple> - you are then restricted to only adding Apple instances to it, but that’s okay, because it is capable of receiving any Fruit.

With these two interfaces, we can make a more versatile fruit bowl. The bowl itself needs to both produce and consume its generic type, so it can neither be covariant nor contravariant, but it can implement our covariant and contravariant interfaces:

class Bowl<T> : Producer<T>, Consumer<T> {
    private val items = mutableListOf<T>()
    override fun get(): T = items.removeAt(items.size - 1)
    override fun add(item: T) { items.add(item) }
}

Now, you can treat a bowl of T as a producer of any superclass of T, and as a consumer of any subclass of T:

val p: Producer<Fruit> = Bowl<Apple>()
val c: Consumer<Apple> = Bowl<Fruit>()

Variance directions

If the parameters or return types of the members of a variant type are themselves variant, it gets a bit complicated. Function types in parameters and return types make it even more challenging. If you’re wondering whether it’s safe to use a variant type parameter T in a particular position, ask yourself:

These considerations lead to the following rules. A covariant type parameter T (which the user of an object might think is Fruit, while the object in reality is tied to Apple) may be used as:

A contravariant type parameter may be used in the converse situations. It is left as an exercise to the reader to figure out the justifications for why these member signatures are legal:

Type projections (use-site covariance and contravariance)

If you’re using a generic class whose type parameters haven’t been declared in a variant way (either because its authors didn’t think of it, or because the type parameters can’t have either variance kind because they are used both as parameter types and return types), you can still use it in a variant way thanks to type projection. The term “projection” refers to the fact that when you do this, you might restrict yourself to using only some of its members - so you’re in a sense only seeing a partial, or “projected” version of the class. Let’s look again at our Bowl class, but without the variant interfaces this time:

class Bowl<T> {
    private val items = mutableListOf<T>()
    fun get(): T = items.removeAt(items.size - 1)
    fun add(item: T) { items.add(item) }
}

Because T is used as a parameter type, it can’t be covariant, and because it’s used as a return type, it can’t be contravariant. But if we only want to use the get() function, we can project it covariantly with out:

fun <T> moveCovariantly(from: Bowl<out T>, to: Bowl<T>) {
    to.add(from.get())
}

Here, we’re saying that the type parameter of from must be a subtype of the type parameter of to. This function will accept e.g. a Bowl<Apple> as from and Bowl<Fruit> as to. The price we’re paying for using the out projection is that we can’t call add() on from(), since we don’t know its true type parameter and we would therefore risk adding incompatible fruits to it.

We could do a similar thing with contravariant projection by using in:

fun <T> moveContravariantly(from: Bowl<T>, to: Bowl<in T>) {
    to.add(from.get())
}

Now, the type parameter of to must be a supertype of that of from. This time, we’re losing the ability to call get() on to.

The same type parameter can be used in both covariant and contravariant projections (because it’s the generic classes that are being projected, not the type parameter):

fun <T> moveContravariantly(from: Bowl<out T>, to: Bowl<in T>) {
    to.add(from.get())
}

While doing so was not useful in this particular example, one could get interesting effects by adding an unprojected parameter type via: Bowl<T>, in which case the generic type parameter of via would be forced to be “in-between” those of from and to.

If you don’t have any idea (or don’t care) what the generic type might be, you can use a star-projection:

fun printSize(items: List<*>) = println(items.size)

When using a generic type where you have star-projected one or more of its type parameters, you can:

Reified type parameters

Sadly, Kotlin has inherited Java’s limitation on generics: they are strictly a compile-time concept - the generic type information is erased at runtime. Therefore, you can not say T() to construct a new instance of a generic type; you can not at runtime check if an object is an instance of a generic type parameter; and if you try to cast between generic types, the compiler can’t guarantee the correctness of it.

Luckily, Kotlin has got reified type parameters, which alleviates some of these problems. By writing reified in front of a generic type parameter, it does become available at runtime, and you’ll get to write T::class to get the class metadata. You can only do this in inline functions (because an inline function will be compiled into its callsite, where the type information is available at runtime), but it still goes a long way. For example, you can make an inline wrapper function for a big function that has got a less elegant signature.

In the example below, we assume that there is a DbModel base class, and that every subclass has got a parameterless primary constructor. In the inline function, T is reified, so we can get the class metadata. We pass this to the function that does the real work of talking to the database.

inline fun <reified T : DbModel> loadFromDb(id: String): T =
    loadFromDb(T::class, id)

fun <T : DbModel> loadFromDb(cls: KClass<T>, id: String): T {
    val entity = cls.primaryConstructor!!.call()
    val tableName = cls.simpleName
    // DB magic goes here - load from table `tableName`,
    // and use the data to populate `entity`
    // (possibly via `memberProperties`)
    return entity
}

Now, you can say loadFromDb<Exercise>("x01234567") to load an object from the Exercise database table.

Extension functions/properties

Since you can’t modify built-in or third-party classes, you can’t directly add functions or properties to them. If you can achieve what you want by only using the public members of a class, you can of course just write a function that takes an instance of the class as a parameter - but sometimes, you’d really like to be able to say x.foo(y) instead of foo(x, y), especially if you want to make a chain of such calls or property lookups: x.foo(y).bar().baz instead of getBaz(bar(foo(x, y))).

There is a nice piece of syntactic sugar that lets you do this: extension functions and extension properties. They look like regular member functions/properties, but they are defined outside of any class - yet they reference the class name and can use this. However, they can only use visible members of the class (typically just the public ones). Behind the scenes, they get compiled down to regular functions that take the target instance as a parameter.

For example, if you work a lot with bytes, you might want to easily get an unsigned byte in the range 0 through 255 instead of the default -128 through 127 (the result will have to be in the form of a Short/Int/Long, though). Byte is a built-in class that you can’t modify, but you can define this extension function:

fun Byte.toUnsigned(): Int {
    return if (this < 0) this + 256 else this.toInt()
}

Now, you can do:

val x: Byte = -1
println(x.toUnsigned()) // Prints 255

If you’d rather do x.unsigned, you can define an extension property:

val Byte.unsigned: Int
    get() = if (this < 0) this + 256 else this.toInt()

Keep in mind that this is just syntactic sugar - you’re not actually modifying the class or its instances. Therefore, you have to import an extension function/property wherever you want to use it (since it isn’t carried along with the instances of the class). For the same reason, you can not override extension members - you can reimplement them for subtypes, but the resolution happens at compile-time based on the static type of the expression you’re invoking it on. So if you declare an extension function for Vehicle, and one with the same name and signature for its subclass Car, and you do the following, it’s the extension function on Vehicle that will be called, even though v is really a Car:

fun foo(v: Vehicle) = v.extension()
val x = foo(Car())

There are a lot of built-in extension functions/properties in Kotlin - for example, map(), filter(), and the rest of the framework for processing collections in a functional manner is built using extension functions.

Member references and reflection

Property references

Consider this class:

class Person(val name: String, var age: Int) {
    fun present() = "I'm $name, and I'm $age years old"
    fun greet(other: String) = "Hi, $other, I'm $name"
}

You can get reference to its name property like this:

val prop = Person::name

The result is an object which represents a reference to the property (the “Platonic ideal” property, not a property on a particular instance). There’s a type hierarchy for property objects: the base interface is KProperty, which lets you get metadata about the property, such as its name and type. If you want to use the property object to read or modify the property’s value in an object, you need to use a subinterface that specifies what kind of property it is. Immutable properties typically are KProperty1<R, V>, and mutable properties typically are KMutableProperty1<R, V>. Both of these are generic interfaces, with R being the receiver type (the type on which the property is declared, in this case Person) and V being the type of the property’s value.

Given an R instance, KProperty1<R, V> will let you read the value that the property has in that instance by calling get(), and KMutableProperty1<R, V> will also let you change the property value in the instance by calling set(). Using this, we can start writing functions that manipulate properties without knowing in advance which property (or which class) they are going to deal with:

fun <T> printProperty(instance: T, prop: KProperty1<T, *>) {
    println("${prop.name} = ${prop.get(instance)}")
}

fun <T> incrementProperty(
    instance: T, prop: KMutableProperty1<T, Int>
) {
    val value = prop.get(instance)
    prop.set(instance, value + 1)
}

val person = Person("Lisa", 23)
printProperty(person, Person::name)
incrementProperty(person, Person::age)

You can also get a reference to a top-level property by just prefixing the property name with :: (e.g. ::foo), and its type will be KProperty0<V> or KMutableProperty0<V>.

Function references

Functions act similarly to properties, but can be referenced as two different kinds of types.

If you want to look at the metadata of a function (e.g. its name), use KFunction<V> or one of its subinterfaces, where V is the function’s return type. Here’s a basic example:

val person = Person("Lisa", 32)
val g: KFunction<String> = Person::greet
println(g.name)
println(g.call(person, "Anne"))

Invoking call() on a function object will call the function. If it is a member function, the first parameter must be the receiver (the object on which the function is to be invoked, in this case person), and the remaining parameters must be the ordinary function parameters (in this case "Anne").

Since the parameter types are not encoded as generic type parameters in KFunction<V>, you won’t get compile-time type validation of the parameters you pass. In order to encode the parameter types, use one of the subinterfaces KFunction1<A, V>, KFunction2<A, B, V>, KFunction3<A, B, C, V>, and so on, depending on how many parameters the function has got. Keep in mind that if you are referencing a member function, the first generic type parameter is the receiver type. For example, KFunction3<A, B, C, V> may reference either an ordinary function that takes A, B, and C as parameters and returns V, or it may reference a member function on A that takes B and C as parameters and returns V. When you use any of these types, you can call the function through its reference as if the reference were a function, e.g. function(a, b), and this call will be type-safe.

You can also reference a member property directly on an object, in which case you get a member function reference that is already bound to its receiver, so that you don’t need the receiver type in the signature. Here’s an example of both approaches:

fun <A, V> callAndPrintOneParam(function: KFunction1<A, V>, a: A): V {
    val result = function(a)
    println("${function.name}($a) = $result")
    return result
}

fun <A, B, V> callAndPrintTwoParam(function: KFunction2<A, B, V>, a: A, b: B): V {
    val result = function(a, b)
    println("${function.name}($a, $b) = $result")
    return result
}

val p = Person("Lisa", 32)
callAndPrintOneParam(p::greet, "Alice")
callAndPrintTwoParam(Person::greet, person, "Lisa")

If you only want to call the function and don’t care about the metadata, use a function type, e.g. (A, B) -> V for an ordinary function reference or a bound member function reference, or A.(B, C) -> V for an unbound member function reference on A. Note that KFunction<V> and its subinterfaces are only available for declared functions (obtained either by explicitly referencing it in the code, or through reflection, as shown later) - only function types are available for function literals (lambda expressions or anonymous functions).

You can get a reference to an top-level function by prefixing the function name with :: (e.g. ::foo).

Obtaining member references from a class reference

While it is possible in Kotlin to dynamically create new classes at runtime or to add members to a class, it’s tricky and slow, and generally discouraged. However, it is easy to dynamically inspect an object to see e.g. what properties and functions it contains and which annotations exist on them. This is called reflection, and it’s not very performant, so avoid it unless you really need it.

Kotlin has got its own reflection library (kotlin-reflect.jar must be included in your build). When targeting the JVM, you can also use the Java reflection facilities. Note that the Kotlin reflection isn’t quite feature-complete yet - in particular, you can’t use it to inspect built-in classes like String.

Warning: using reflection is usually the wrong way to solve problems in Kotlin! In particular, if you have several classes that all have some common properties/functions and you want to write a function that can take an instance of any of those classes and use those properties, the correct approach is to define an interface with the common properties/functions and make all the relevant classes implement it; the function can then take that interface as a parameter. If you don’t control those classes, you can use the Adapter pattern and write wrapper classes that implement the interface - this is very easy thanks to Kotlin’s delegation feature. You can also get a lot of leverage out of using generics in clever ways.

Appending ::class to a class name will give you a KClass<C> metadata object for that class. The generic type parameter C is the class itself, so you can use KClass<*> if you’re writing a function that can work with metadata for any class, or you can make a generic function with a type parameter T and parameter type KClass<T>. From this, you can obtain references to the members of the class. The most interesting properties on KClass are probably primaryConstructor, constructors, memberProperties, declaredMemberProperties, memberFunctions, and declaredMemberFunctions. The difference between e.g. memberProperties and declaredMemberProperties is that the former includes inherited properties, while the latter only includes the properties that have been declared in the class’ own body.

In this example, using Person and callAndPrintTwoParam() from the previous section, we locate a member function reference by name and call it:

val f = Person::class.memberFunctions.single { it.name == "greet" } as KFunction2<Person, String, String>
callAndPrintTwoParam(f, person, "Lisa")

The signature of greet() is KFunction2<Person, String, String> because it’s a function on Person that takes a String and returns a String.

Constructor references are effectively factory functions for creating new instances of a class, which might come in handy:

val ctor = Person::class.primaryConstructor!! as (String, Int) -> Person
val newPerson = ctor("Karen", 45)

Java-style reflection

If you’re targeting the JVM platform, you can also use Java’s reflection system directly. In this example, we grab a function reference from an object’s class by specifying the function’s name as a string (if the function takes parameters, you also need to specify their types), and then we call it. Note that we didn’t mention String anywhere - this technique works without knowing what the object’s class is, but it will raise an exception if the object’s class doesn’t have the requested function. However, Java-style function references do not have type information, so you won’t get verification of the parameter types, and you must cast the return value:

val s = "Hello world"
val length = s.javaClass.getMethod("length")
val x = length.invoke(s) as Int

If you don’t have an instance of the class, you can get the class metadata with String::class.java (but you can’t invoke any of its members until you have an instance).

If you need to look up the class dynamically as well, you can use Class.forName() and supply the fully-qualified name of the class.

Annotations

While Kotlin annotations look like Python decorators, they are far less flexible: they can generally only be used for metadata. They are pure data-containing classes, and do not contain any executable code. Some built-in annotations have an effect on the compilation process (such as @JvmStatic), but custom annotations are only useful for providing metadata that can be inspected at runtime by the reflection system. We won’t delve deeply into annotations here, but here is an example. The annotations on the annotation declaration itself specify what constructs the annotation may be applied to and whether it is available for runtime inspection.

enum class TestSizes { SMALL, MEDIUM, LARGE }

@Target(AnnotationTarget.CLASS)
@Retention(AnnotationRetention.RUNTIME)
annotation class TestSize(val size: TestSizes)

@TestSize(TestSizes.SMALL)
class Tests { ... }

fun getTestSize(cls: KClass<*>): TestSizes? =
    cls.findAnnotation<TestSize>()?.size

println(getTestSize(Tests::class))

File I/O

Kotlin has inherited Java’s fidgety (but very flexible) way of doing I/O, but with some simplifying extra features. We won’t get into all of it here, so for starters, this is how to iterate through all the lines of a file (you’ll need import java.io.File):

File("data.txt").forEachLine { println(it) }

The default encoding is UTF-8, but you can specify it if you need something else:

File("data.txt").forEachLine(Charsets.UTF_16) { println(it) }

Note that the trailing newline of each line is stripped. You can also call readLines() on a file object to get a list of all the lines, or useLines() to supply a function that will be called on every line. If you simply want the entire file contents as one string or byte array, call readText() or readBytes(), respectively.

Note that while File() does create a “file object”, it doesn’t actually open the file - the file object is just a reference to the file path; opening the file is a separate action. The preceding functions open and close the file automatically, whereas other functions separately open and close the file. For example, if you’re parsing binary data and you don’t want to read the entire file at once, you must create an input stream (for binary data) or an input stream reader (for strings) - the example below will read 16 bytes:

val stream = File("data.txt").inputStream()
val bytes = ByteArray(16)
stream.read(bytes)
stream.close()
println(bytes)

It’s important to close a stream when you’re done with it; otherwise, your program will leak a file handle. See the next section for how do do this nicely.

If you’ve got one string that you want to write to a file, overwriting the existing contents if the file already exists, do this (again, UTF-8 is the default encoding):

File("data.txt").writeText("Hello world!")

If you want to write strings gradually, you need to create an OutputStreamWriter by calling writer() on the file object. You can write binary data to a file by calling outputStream() on a file object and use the resulting OutputStream to write bytes.

If you need a fancier way of reading or writing file data, you have access to the full Java suite of I/O classes - in particular, Scanner, which can parse numbers and other data types from files or other streams, and BufferedReader (which is good for efficient reading of large amounts of data), which you can obtain by calling bufferedReader() on a file or stream. See any Java tutorial for how to use these.

Scoped resource usage

Kotlin does not have Python’s resource managers or Java’s try-with-resources, but thanks to extension functions, there’s use:

File("/home/aasmund/test.txt").inputStream().use {
     val bytes = it.readBytes()
     println(bytes.size)
}

use can be invoked on anything that implements the Closeable interface, and when the use block ends (whether normally or due to an exception), close() will be called on the object upon which you invoked use. If an exception is raised within the block or by close(), it will bubble out of use. If both the block and close() raise, it’s the exception from the block that will bubble out.

Thus, you can create something resource manager-like by creating a class that implements Closeable, does its setup work in init, and does its cleanup work in close().

In case you’re wondering about how use, which is a function, can just be followed by a block like that, see the section on DSL support.

Documentation

Kotlin’s documentation syntax is called KDoc. A KDoc block is placed above the construct it describes, and begins with /** and ends with */ (possibly on one line; if not, each intermediate lines should start with an aligned asterisk). The first block of text is the summary; then, you can use block tags to provide information about specific parts of the construct. Some block tags are @param for function parameters and generic type parameters, and @return for the return value. You can link to identifiers inside brackets. All the text outside of links and block tag names is in Markdown format.

/**
 * Squares a number.
 *
 * @param number Any [Double] number whose absolute value is
 * less than or equal to the square root of [Double.MAX_VALUE].
 * @return A nonnegative number: the result of multiplying [number] with itself.
 */
fun square(number: Double) = number * number

Package-level documentation can be provided in a separate Markdown file.

Unlike docstrings, KDoc blocks are not available to the program at runtime.

You can generate separate documentation files in HTML format from KDoc by using a tool called Dokka.


This material was written by Aasmund Eldhuset; it is owned by Khan Academy and licensed for use under CC BY-NC-SA 3.0 US.