我可能有一个像下面这样的数组:

[1, 4, 2, 2, 6, 24, 15, 2, 60, 15, 6]

或者,实际上,任何类似类型的数据部分的序列。我要做的是确保每个相同的元素只有一个。例如,上面的数组将变成:

[1, 4, 2, 6, 24, 15, 60]

请注意,删除了2、6和15的重复项,以确保每个相同的元素中只有一个。Swift是否提供了一种容易做到这一点的方法,还是我必须自己做?


当前回答

受https://www.swiftbysundell.com/posts/the-power-of-key-paths-in-swift的启发,我们可以声明一个更强大的工具,它能够过滤任何keyPath上的唯一性。感谢Alexander关于复杂性的各种答案的评论,下面的解决方案应该是最优的。

Non-mutating解决方案

我们扩展了一个函数,它可以过滤任意keyPath上的唯一性:

extension RangeReplaceableCollection {
    /// Returns a collection containing, in order, the first instances of
    /// elements of the sequence that compare equally for the keyPath.
    func unique<T: Hashable>(for keyPath: KeyPath<Element, T>) -> Self {
        var unique = Set<T>()
        return filter { unique.insert($0[keyPath: keyPath]).inserted }
    }
}

注意:在你的对象不符合RangeReplaceableCollection,但符合Sequence的情况下,你可以有这个额外的扩展,但返回类型将始终是一个数组:

extension Sequence {
    /// Returns an array containing, in order, the first instances of
    /// elements of the sequence that compare equally for the keyPath.
    func unique<T: Hashable>(for keyPath: KeyPath<Element, T>) -> [Element] {
        var unique = Set<T>()
        return filter { unique.insert($0[keyPath: keyPath]).inserted }
    }
}

使用

如果我们想要元素本身的唯一性,如问题中所示,我们使用keyPath \.self:

let a = [1, 4, 2, 2, 6, 24, 15, 2, 60, 15, 6]
let b = a.unique(for: \.self)
/* b is [1, 4, 2, 6, 24, 15, 60] */

如果我们想要其他东西的唯一性(比如对象集合的id),那么我们使用我们选择的keyPath:

let a = [CGPoint(x: 1, y: 1), CGPoint(x: 2, y: 1), CGPoint(x: 1, y: 2)]
let b = a.unique(for: \.y)
/* b is [{x 1 y 1}, {x 1 y 2}] */

变异的解决方案

我们扩展了一个变异函数,它能够过滤任意keyPath上的唯一性:

extension RangeReplaceableCollection {
    /// Keeps only, in order, the first instances of
    /// elements of the collection that compare equally for the keyPath.
    mutating func uniqueInPlace<T: Hashable>(for keyPath: KeyPath<Element, T>) {
        var unique = Set<T>()
        removeAll { !unique.insert($0[keyPath: keyPath]).inserted }
    }
}

使用

如果我们想要元素本身的唯一性,如问题中所示,我们使用keyPath \.self:

var a = [1, 4, 2, 2, 6, 24, 15, 2, 60, 15, 6]
a.uniqueInPlace(for: \.self)
/* a is [1, 4, 2, 6, 24, 15, 60] */

如果我们想要其他东西的唯一性(比如对象集合的id),那么我们使用我们选择的keyPath:

var a = [CGPoint(x: 1, y: 1), CGPoint(x: 2, y: 1), CGPoint(x: 1, y: 2)]
a.uniqueInPlace(for: \.y)
/* a is [{x 1 y 1}, {x 1 y 2}] */

其他回答

下面是SequenceType上的一个类别,它保留了数组的原始顺序,但使用Set来进行contains查找,以避免数组的contains(_:)方法上的O(n)代价。

public extension Sequence where Element: Hashable {

    /// Return the sequence with all duplicates removed.
    ///
    /// i.e. `[ 1, 2, 3, 1, 2 ].uniqued() == [ 1, 2, 3 ]`
    ///
    /// - note: Taken from stackoverflow.com/a/46354989/3141234, as 
    ///         per @Alexander's comment.
    func uniqued() -> [Element] {
        var seen = Set<Element>()
        return self.filter { seen.insert($0).inserted }
    }
}

如果你不是Hashable或Equatable,你可以传入一个谓词来进行相等性检查:

extension Sequence {

    /// Return the sequence with all duplicates removed.
    ///
    /// Duplicate, in this case, is defined as returning `true` from `comparator`.
    ///
    /// - note: Taken from stackoverflow.com/a/46354989/3141234
    func uniqued(comparator: @escaping (Element, Element) throws -> Bool) rethrows -> [Element] {
        var buffer: [Element] = []

        for element in self {
            // If element is already in buffer, skip to the next element
            if try buffer.contains(where: { try comparator(element, $0) }) {
                continue
            }

            buffer.append(element)
        }

        return buffer
    }
}

现在,如果你没有Hashable,但是是Equatable,你可以使用这个方法:

extension Sequence where Element: Equatable {

    /// Return the sequence with all duplicates removed.
    ///
    /// i.e. `[ 1, 2, 3, 1, 2 ].uniqued() == [ 1, 2, 3 ]`
    ///
    /// - note: Taken from stackoverflow.com/a/46354989/3141234
    func uniqued() -> [Element] {
        return self.uniqued(comparator: ==)
    }
}

最后,你可以添加一个unique的关键路径版本,如下所示:

extension Sequence {

    /// Returns the sequence with duplicate elements removed, performing the comparison using the property at
    /// the supplied keypath.
    ///
    /// i.e.
    ///
    /// ```
    /// [
    ///   MyStruct(value: "Hello"),
    ///   MyStruct(value: "Hello"),
    ///   MyStruct(value: "World")
    ///  ].uniqued(\.value)
    /// ```
    /// would result in
    ///
    /// ```
    /// [
    ///   MyStruct(value: "Hello"),
    ///   MyStruct(value: "World")
    /// ]
    /// ```
    ///
    /// - note: Taken from stackoverflow.com/a/46354989/3141234
    ///
    func uniqued<T: Equatable>(_ keyPath: KeyPath<Element, T>) -> [Element] {
        self.uniqued { $0[keyPath: keyPath] == $1[keyPath: keyPath] }
    }
}

你可以把这两个都放在你的应用程序中,Swift会根据你的序列的迭代器选择正确的一个。元素类型。


对于El Capitan,您可以扩展此方法以包括多个键盘,如下所示:

    /// Returns the sequence with duplicate elements removed, performing the comparison using the property at
    /// the supplied keypaths.
    ///
    /// i.e.
    ///
    /// ```
    /// [
    ///   MyStruct(value1: "Hello", value2: "Paula"),
    ///   MyStruct(value1: "Hello", value2: "Paula"),
    ///   MyStruct(value1: "Hello", value2: "Bean"),
    ///   MyStruct(value1: "World", value2: "Sigh")
    ///  ].uniqued(\.value1, \.value2)
    /// ```
    /// would result in
    ///
    /// ```
    /// [
    ///   MyStruct(value1: "Hello", value2: "Paula"),
    ///   MyStruct(value1: "Hello", value2: "Bean"),
    ///   MyStruct(value1: "World", value2: "Sigh")
    /// ]
    /// ```
    ///
    /// - note: Taken from stackoverflow.com/a/46354989/3141234
    ///
    func uniqued<T: Equatable, U: Equatable>(_ keyPath1: KeyPath<Element, T>, _ keyPath2: KeyPath<Element, U>) -> [Element] {
        self.uniqued {
            $0[keyPath: keyPath1] == $1[keyPath: keyPath1] && $0[keyPath: keyPath2] == $1[keyPath: keyPath2]
        }
    }

但是(恕我直言)你最好把你自己的block传递给self.unique。

让我提出一个类似于斯科特·加德纳的答案,但使用了更简洁的reduce语法。 此解决方案从自定义对象数组中删除重复项(保持初始顺序)

// Custom Struct. Can be also class. 
// Need to be `equitable` in order to use `contains` method below
struct CustomStruct : Equatable {
      let name: String
      let lastName : String
    }

// conform to Equatable protocol. feel free to change the logic of "equality"
func ==(lhs: CustomStruct, rhs: CustomStruct) -> Bool {
  return (lhs.name == rhs.name && lhs.lastName == rhs.lastName)
}

let categories = [CustomStruct(name: "name1", lastName: "lastName1"),
                  CustomStruct(name: "name2", lastName: "lastName1"),
                  CustomStruct(name: "name1", lastName: "lastName1")]
print(categories.count) // prints 3

// remove duplicates (and keep initial order of elements)
let uniq1 : [CustomStruct] = categories.reduce([]) { $0.contains($1) ? $0 : $0 + [$1] }
print(uniq1.count) // prints 2 - third element has removed

如果你想知道这个约简魔法是如何工作的,这里是完全相同的,只是使用了更扩展的约简语法

let uniq2 : [CustomStruct] = categories.reduce([]) { (result, category) in
  var newResult = result
  if (newResult.contains(category)) {}
  else {
    newResult.append(category)
  }
  return newResult
}
uniq2.count // prints 2 - third element has removed

你可以简单地复制粘贴这段代码到Swift Playground中。

另一种(如果不是最优的)解决方案是使用不可变类型而不是变量:

func deleteDuplicates<S: ExtensibleCollectionType where S.Generator.Element: Equatable>(seq:S)-> S {
    let s = reduce(seq, S()){
        ac, x in contains(ac,x) ? ac : ac + [x]
    }
    return s
}

包括对比Jean-Pillippe的命令式方法和函数式方法。

作为奖励,这个函数不仅可以处理数组,还可以处理字符串!

编辑:这个答案是在2014年为Swift 1.0编写的(在Set在Swift中可用之前)。它不需要Hashable一致性,并且在二次时间内运行。

这里有很多答案,但我错过了这个简单的扩展,适合Swift 2及以上:

extension Array where Element:Equatable {
    func removeDuplicates() -> [Element] {
        var result = [Element]()

        for value in self {
            if result.contains(value) == false {
                result.append(value)
            }
        }

        return result
    }
}

这非常简单。可以这样调用:

let arrayOfInts = [2, 2, 4, 4]
print(arrayOfInts.removeDuplicates()) // Prints: [2, 4]

基于属性的过滤

要根据属性筛选数组,你可以使用这个方法:

extension Array {

    func filterDuplicates(@noescape includeElement: (lhs:Element, rhs:Element) -> Bool) -> [Element]{
        var results = [Element]()

        forEach { (element) in
            let existingElements = results.filter {
                return includeElement(lhs: element, rhs: $0)
            }
            if existingElements.count == 0 {
                results.append(element)
            }
        }

        return results
    }
}

你可以这样调用它:

let filteredElements = myElements.filterDuplicates { $0.PropertyOne == $1.PropertyOne && $0.PropertyTwo == $1.PropertyTwo }

你可以自己卷,比如这样:

func unique<S : Sequence, T : Hashable>(source: S) -> [T] where S.Iterator.Element == T {
    var buffer = [T]()
    var added = Set<T>()
    for elem in source {
        if !added.contains(elem) {
            buffer.append(elem)
            added.insert(elem)
        }
    }
    return buffer
}

let vals = [1, 4, 2, 2, 6, 24, 15, 2, 60, 15, 6]
let uniqueVals = uniq(vals) // [1, 4, 2, 6, 24, 15, 60]

作为Array的扩展:

extension Array where Element: Hashable {
    func uniqued() -> Array {
        var buffer = Array()
        var added = Set<Element>()
        for elem in self {
            if !added.contains(elem) {
                buffer.append(elem)
                added.insert(elem)
            }
        }
        return buffer
    }
}

或者更优雅一点(Swift 4/5):

extension Sequence where Element: Hashable {
    func uniqued() -> [Element] {
        var set = Set<Element>()
        return filter { set.insert($0).inserted }
    }
}

将被使用:

[1,2,4,2,1].uniqued()  // => [1,2,4]