Taking examples from [Python 3 Patterns, Recipes and Idioms](
[To see links please register here]
) we can convert each one to Kotlin using a simple pattern. The Python version of a list comprehension has 3 parts:
1. output expression
2. input list/sequence and variable
3. 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:
<!-- language: python -->
# === 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 ]
Becomes
<!-- language: kotlin -->
// === 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:
<!-- language: python -->
# === 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()`:
<!-- language: kotlin -->
// === KOTLIN
val evenIntsSquared = aList.asSequence().filter { it % 2 == 0 }.map { it * it }
<hr/>
Nested comprehensions in Kotlin are created by just nesting one within the other's `map` lambda. For example, take this sample from [PythonCourse.eu](
[To see links please register here]
) in Python changed slightly to use both a set and a list comprehension:
<!-- language: python -->
# === 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]
Becomes:
<!-- language: kotlin -->
// === 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()`:
<!-- language: kotlin -->
// === 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 }
<hr/>
Another example from the referenced Python page is generating a matrix:
<!-- language: python -->
# === 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]]
And in Kotlin:
<!-- language: kotlin -->
// === 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:
<!-- language: kotlin -->
// === KOTLIN
val matrix2 = Array(3) { row ->
IntArray(3) { col -> if (col == row) 1 else 0 }
}
<hr/>
Another of the examples for set comprehensions is to generate a unique set of properly cased names:
<!-- language: python -->
# === 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'}
Is translated to Kotlin:
<!-- language: kotlin -->
// === 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]
<hr/>
And the example for map comprehension is a bit odd, but can also be implemented in Kotlin. The original:
<!-- language: python -->
# === 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:
<!-- language: kotlin -->
// === 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}
<hr/>
Further reading:
* 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?
](
[To see links please register here]
) for more examples.
* See [Get Factors of Numbers in Kotlin
](
[To see links please register here]
) which shows another example of a Python comprehension versus Kotlin.
* See [Kotlin Extensions Functions for Collections](
[To see links please register here]
) in the API reference guide.