Swift: what is the right way to split up a [String] resulting in a [[String]] with a given subarray size?
In Swift 3/4 this would look like the following:
let numbers = ["1","2","3","4","5","6","7"]
let chunkSize = 2
let chunks = stride(from: 0, to: numbers.count, by: chunkSize).map {
Array(numbers[$0..<min($0 + chunkSize, numbers.count)])
}
// prints as [["1", "2"], ["3", "4"], ["5", "6"], ["7"]]
As an extension to Array:
extension Array {
func chunked(by chunkSize: Int) -> [[Element]] {
return stride(from: 0, to: self.count, by: chunkSize).map {
Array(self[$0..<Swift.min($0 + chunkSize, self.count)])
}
}
}
Or the slightly more verbose, yet more general:
let numbers = ["1","2","3","4","5","6","7"]
let chunkSize = 2
let chunks: [[String]] = stride(from: 0, to: numbers.count, by: chunkSize).map {
let end = numbers.endIndex
let chunkEnd = numbers.index($0, offsetBy: chunkSize, limitedBy: end) ?? end
return Array(numbers[$0..<chunkEnd])
}
This is more general because I am making fewer assumptions about the type of the index into the collection. In the previous implementation I assumed that they could be could be compared and added.
Note that in Swift 3 the functionality of advancing indices has been transferred from the indices themselves to the collection.
With Swift 5, according to your needs, you can choose one of the five following ways in order to solve your problem.
1. Using AnyIterator
in a Collection
extension method
AnyIterator
is a good candidate to iterate over the indices of an object that conforms to Collection
protocol in order to return subsequences of this object. In a Collection
protocol extension, you can declare a chunked(by:)
method with the following implementation:
extension Collection {
func chunked(by distance: Int) -> [[Element]] {
precondition(distance > 0, "distance must be greater than 0") // prevents infinite loop
var index = startIndex
let iterator: AnyIterator<Array<Element>> = AnyIterator({
let newIndex = self.index(index, offsetBy: distance, limitedBy: self.endIndex) ?? self.endIndex
defer { index = newIndex }
let range = index ..< newIndex
return index != self.endIndex ? Array(self[range]) : nil
})
return Array(iterator)
}
}
Usage:
let array = ["1", "2", "3", "4", "5", "6", "7", "8", "9"]
let newArray = array.chunked(by: 2)
print(newArray) // prints: [["1", "2"], ["3", "4"], ["5", "6"], ["7", "8"], ["9"]]
2. Using stride(from:to:by:)
function in an Array
extension method
Array
indices are of type Int
and conform to Strideable
protocol. Therefore, you can use stride(from:to:by:)
and advanced(by:)
with them. In an Array
extension, you can declare a chunked(by:)
method with the following implementation:
extension Array {
func chunked(by distance: Int) -> [[Element]] {
let indicesSequence = stride(from: startIndex, to: endIndex, by: distance)
let array: [[Element]] = indicesSequence.map {
let newIndex = $0.advanced(by: distance) > endIndex ? endIndex : $0.advanced(by: distance)
//let newIndex = self.index($0, offsetBy: distance, limitedBy: self.endIndex) ?? self.endIndex // also works
return Array(self[$0 ..< newIndex])
}
return array
}
}
Usage:
let array = ["1", "2", "3", "4", "5", "6", "7", "8", "9"]
let newArray = array.chunked(by: 2)
print(newArray) // prints: [["1", "2"], ["3", "4"], ["5", "6"], ["7", "8"], ["9"]]
3. Using a recursive approach in an Array
extension method
Based on Nate Cook recursive code, you can declare a chunked(by:)
method in an Array
extension with the following implementation:
extension Array {
func chunked(by distance: Int) -> [[Element]] {
precondition(distance > 0, "distance must be greater than 0") // prevents infinite loop
if self.count <= distance {
return [self]
} else {
let head = [Array(self[0 ..< distance])]
let tail = Array(self[distance ..< self.count])
return head + tail.chunked(by: distance)
}
}
}
Usage:
let array = ["1", "2", "3", "4", "5", "6", "7", "8", "9"]
let newArray = array.chunked(by: 2)
print(newArray) // prints: [["1", "2"], ["3", "4"], ["5", "6"], ["7", "8"], ["9"]]
4. Using a for loop and batches in a Collection
extension method
Chris Eidhof and Florian Kugler show in Swift Talk #33 - Sequence & Iterator (Collections #2) video how to use a simple for loop to fill batches of sequence elements and append them on completion to an array. In a Sequence
extension, you can declare a chunked(by:)
method with the following implementation:
extension Collection {
func chunked(by distance: Int) -> [[Element]] {
var result: [[Element]] = []
var batch: [Element] = []
for element in self {
batch.append(element)
if batch.count == distance {
result.append(batch)
batch = []
}
}
if !batch.isEmpty {
result.append(batch)
}
return result
}
}
Usage:
let array = ["1", "2", "3", "4", "5", "6", "7", "8", "9"]
let newArray = array.chunked(by: 2)
print(newArray) // prints: [["1", "2"], ["3", "4"], ["5", "6"], ["7", "8"], ["9"]]
5. Using a custom struct
that conforms to Sequence
and IteratorProtocol
protocols
If you don't want to create extensions of Sequence
, Collection
or Array
, you can create a custom struct
that conforms to Sequence
and IteratorProtocol
protocols. This struct
should have the following implementation:
struct BatchSequence<T>: Sequence, IteratorProtocol {
private let array: [T]
private let distance: Int
private var index = 0
init(array: [T], distance: Int) {
precondition(distance > 0, "distance must be greater than 0") // prevents infinite loop
self.array = array
self.distance = distance
}
mutating func next() -> [T]? {
guard index < array.endIndex else { return nil }
let newIndex = index.advanced(by: distance) > array.endIndex ? array.endIndex : index.advanced(by: distance)
defer { index = newIndex }
return Array(array[index ..< newIndex])
}
}
Usage:
let array = ["1", "2", "3", "4", "5", "6", "7", "8", "9"]
let batchSequence = BatchSequence(array: array, distance: 2)
let newArray = Array(batchSequence)
print(newArray) // prints: [["1", "2"], ["3", "4"], ["5", "6"], ["7", "8"], ["9"]]
I wouldn't call it beautiful, but here's a method using map
:
let numbers = ["1","2","3","4","5","6","7"]
let splitSize = 2
let chunks = numbers.startIndex.stride(to: numbers.count, by: splitSize).map {
numbers[$0 ..< $0.advancedBy(splitSize, limit: numbers.endIndex)]
}
The stride(to:by:)
method gives you the indices for the first element of each chunk, so you can map those indices to a slice of the source array using advancedBy(distance:limit:)
.
A more "functional" approach would simply be to recurse over the array, like so:
func chunkArray<T>(s: [T], splitSize: Int) -> [[T]] {
if countElements(s) <= splitSize {
return [s]
} else {
return [Array<T>(s[0..<splitSize])] + chunkArray(Array<T>(s[splitSize..<s.count]), splitSize)
}
}