Answered Aug 17, 2019 · 41 votes
Taking examples from Python 3 Patterns, Recipes and Idioms we can convert each one to Kotlin using a simple pattern. The Python version of a list comprehension has 3 parts:
- output expression
- input list/sequence and variable
- optional predicate
These directly correlate to Kotlin functional extensions to collection classes. The input sequence, followed by the optional predicate in a filter lambda, followed by the output expression in a map lambda. So for this Python example:
# === PYTHON-a_list = [1, 2, 3, 4, 5, 6]-# output | var | input | filter/predicate even_ints_squared = [ e*e for e in a_list if e % 2 == 0 ]print(even_ints_squared)# output: [ 4, 16, 36 ]
// === KOTLIN-var aList = listOf(1, 2, 3, 4, 5, 6)-// input | filter | output val evenIntsSquared = aList.filter { it % 2 == 0 }.map { it * it }println(evenIntsSquared)// output: [ 4, 16, 36 ]
Notice that the variable is not needed in the Kotlin version since the implied it variable is used within each lambda. In Python you can turn these into a lazy generator by using the () instead of square brackets:
# === PYTHON-even_ints_squared = ( e**2 for e in a_list if e % 2 == 0 )-
And in Kotlin it is more obviously converted to a lazy sequence by changing the input via a function call asSequence():
// === KOTLIN-val evenIntsSquared = aList.asSequence().filter { it % 2 == 0 }.map { it * it }-
Nested comprehensions in Kotlin are created by just nesting one within the other's map lambda. For example, take this sample from PythonCourse.eu in Python changed slightly to use both a set and a list comprehension:
# === PYTHON-noprimes = {j for i in range(2, 8) for j in range(i*2, 100, i)}primes = [x for x in range(2, 100) if x not in noprimes]print(primes)# output: [2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97]
// === KOTLIN-val nonprimes = (2..7).flatMap { (it*2..99).step(it).toList() }.toSet()val primes = (2..99).filterNot { it in nonprimes }print(primes) // output: [2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97]
Notice that the nested comprehension produces a list of lists which is converted to a flat list using flatMap() and then converted to a set using toSet(). Also, Kotlin ranges are inclusive, whereas a Python range is exclusive so you will see the numbers are slightly different in the ranges.
You can also use a sequence generator with co-routines in Kotlin to yield the values without needing the call to flatMap() or flatten():
// === KOTLIN-val nonprimes = sequence { (2..7).forEach { (it*2..99).step(it).forEach { value -> yield(value) } }}.toSet()val primes = (2..99).filterNot { it in nonprimes }
Another example from the referenced Python page is generating a matrix:
# === PYTHON-matrix = [ [ 1 if item_idx == row_idx else 0 for item_idx in range(0, 3) ] for row_idx in range(0, 3) ]print(matrix)# [[1, 0, 0], # [0, 1, 0], # [0, 0, 1]]
// === KOTLIN-val matrix = (0..2).map { row -> (0..2).map { col -> if (col == row) 1 else 0 }}println(matrix)// [[1, 0, 0], // [0, 1, 0], // [0, 0, 1]]
Or in Kotlin instead of lists, you could also generate arrays:
// === KOTLIN-val matrix2 = Array(3) { row -> IntArray(3) { col -> if (col == row) 1 else 0 } }
Another of the examples for set comprehensions is to generate a unique set of properly cased names:
# === PYTHON-names = [ 'Bob', 'JOHN', 'alice', 'bob', 'ALICE', 'J', 'Bob' ]-fixedNames = { name[0].upper() + name[1:].lower() for name in names if len(name) > 1 }print(fixedNames)# output: {'Bob', 'Alice', 'John'}
// === KOTLIN-val names = listOf( "Bob", "JOHN", "alice", "bob", "ALICE", "J", "Bob" )-val fixedNames = names.filter { it.length > 1 } .map { it.take(1).toUpperCase() + it.drop(1).toLowerCase() } .toSet()println(fixedNames)// output: [Bob, John, Alice]
And the example for map comprehension is a bit odd, but can also be implemented in Kotlin. The original:
# === PYTHON-mcase = {'a':10, 'b': 34, 'A': 7, 'Z':3}-mcase_frequency = { k.lower() : mcase.get(k.lower(), 0) + mcase.get(k.upper(), 0) for k in mcase.keys() }print(mcase_frequency)# output: {'a': 17, 'z': 3, 'b': 34}
And the converted, which is written to be a bit more "wordy" here to make it clearer what is happening:
// === KOTLIN-val mcase = mapOf("a" to 10, "b" to 34, "A" to 7, "Z" to 3)-val mcaseFrequency = mcase.map { (key, _) -> val newKey = key.toLowerCase() val newValue = mcase.getOrDefault(key.toLowerCase(), 0) + mcase.getOrDefault(key.toUpperCase(), 0) newKey to newValue}.toMap()print(mcaseFrequency)// output: {a=17, b=34, z=3}
- Kotlin adds more power than list/set/map comprehensions because of its extensive functional transforms that you can make to these collection types. See What Java 8 Stream.collect equivalents are available in the standard Kotlin library? for more examples.
- See Get Factors of Numbers in Kotlin which shows another example of a Python comprehension versus Kotlin.
- See Kotlin Extensions Functions for Collections in the API reference guide.