是否可以从PriorityQueue中删除元素？

Question

是否可以从PriorityQueue中删除元素？

文档：
http://www.scala-lang.org/api/current/index.html#scala.collection.mutable.PriorityQueue
http://www.scala-lang.org/api/current/index.html#scala.collection.Iterator

我有一个PQ w各种双值（一些重复） - 我用它作为堆来跟踪流媒体环境中的滚动中位数。我想从PQ中删除值但无法弄清楚如何。
我试图使用迭代器来查找PQ的一个元素并放在那里，但它没有用。我想知道它是否可能？

val maxHeapLeft= new mutable.PriorityQueue[Double]()(Ordering[Double])
maxHeapLeft.enqueue(5)
maxHeapLeft.enqueue(55)
maxHeapLeft.enqueue(25)
maxHeapLeft.enqueue(15)
maxHeapLeft.enqueue(15)
val it= maxHeapLeft.iterator
var p1=it.next
p1=it.next

println("size before " +maxHeapLeft.size)
it.drop(1)
println("size AFTER " +maxHeapLeft.size)

PQ的大小不会改变。

编辑1：到目前为止，我使用maxHeapLeft= new mutable.PriorityQueue[Double]()(Ordering[Double]) ++ (maxHeapLeft.toList diff List(15))从PQ中移除了15。当然，太可怕了。

编辑2：自定义优先级队列失败的测试用例（对于@Nate）：

 "PQ" should "produce correct values " in {
    val testOperations = List[String]("8114.0", "9233.0", "dequeue", "10176.0", "10136.0", "dequeue", "10041.0", "9900.0", "10787.0", "10476.0", "10439.0", "dequeue", "10722.0", "9900.0", "11028.0", "10764.0", "dequeue", "10698.0", "10374.0", "dequeue", "-10176.0", "10198.0", "-10136.0", "11478.0", "10930.0", "dequeue", "10881.0", "dequeue", "10555.0", "dequeue", "-10787.0", "10439.0", "-10476.0", "11596.0", "-10439.0", "10757.0", "-10722.0", "10493.0", "10551.0", "dequeue", "-11028.0", "10493.0", "-10764.0", "11892.0", "-10698.0", "11276.0", "10917.0", "dequeue", "15855.0", "dequeue", "12008.0", "dequeue")
    val customPQ= new PriorityQueue[Double]()(Ordering[Double].reverse) //cread min heap

    for (el <-testOperations){
      el match {
        case dequeue if el=="dequeue" => customPQ.dequeue()
        case remove if remove.toDouble < 0 => customPQ -= (-1*remove.toDouble )
        case add => customPQ.enqueue(add.toDouble )
      }
    }

    println(customPQ.head + "==" + customPQ.min)
    println(customPQ)
  }

测试输出：
10881.0 == 10757.0
PriorityQueue（10881.0,10917.0,11596.0,10930.0,11276.0,11892.0,12008.0,11478.0,10757.0,15855.0）

Answer 1

根据文档，您只能按clear和dequeue删除元素。

也许您对TreeMultiset获得所需功能的成本增加感到满意。

如果你想删除堆中的特定值，那么你可以从source开始自己滚动。

修改

提供O(n)删除的

Here is an updated version of PriorityQueue。以下是相关的已添加代码段：

def -=(elem: A): this.type = {
  var k: Int = find(elem)
  resarr.p_size0 = resarr.p_size0 - 1
  resarr.p_swap(k, resarr.p_size0)
  fixUp(resarr.p_array, k)
  fixDown(resarr.p_array, k, resarr.p_size0 - 1)
  this
}

protected def find(elem: A): Int = {
  var k: Int = 1
  while (k < resarr.length) {
    if (resarr.p_array(k) == elem) {
      return k
    }
    k += 1
  }
  throw new NoSuchElementException("element does not exist in heap")
}

如果他/她希望移除MultiMap，我会向读者/ OP添加O(lg n)作为练习。 （提示：您需要更新修改resarr数组的所有方法。）

编辑2：

在本地运行：

$ scalac -version 
Scala compiler version
2.11.2 -- Copyright 2002-2013, LAMP/EPFL 

$ md5 PriorityQueue.scala  
MD5 (PriorityQueue.scala) = 3913496441f83bcdeda2249ec2a6b574 

$ scalac PriorityQueue.scala  

$ scala Test 
size before 4 
size after 3

Answer 2

PriorityQueue执行任务的正确性。如果您需要具有唯一值的类似API，请使用SortedSet。

Answer 3

您始终可以过滤掉单个元素（在非并发代码中）：

var filtered = false
pq.filter(x => !filtered && { val ok = p(x); if (!ok) filtered = true; ok })

也就是说，这是每次重建整个队列（保存一个元素），这对你来说可能不够快。