Open In App
Related Articles

Priority Queue in Python

Improve
Improve
Improve
Like Article
Like
Save Article
Save
Report issue
Report

Priority Queues are abstract data structures where each data/value in the queue has a certain priority. For example, In airlines, baggage with the title “Business” or “First-class” arrives earlier than the rest.

Priority Queue is an extension of the queue with the following properties.

  1. An element with high priority is dequeued before an element with low priority.
  2. If two elements have the same priority, they are served according to their order in the queue.
    Various applications of the Priority queue in Computer Science are:
    Job Scheduling algorithms, CPU and Disk Scheduling, managing resources that are shared among different processes, etc.

Key differences between Priority Queue and Queue:

  1. In Queue, the oldest element is dequeued first. While, in Priority Queue, an element based on the highest priority is dequeued.
  2. When elements are popped out of a priority queue the result obtained is either sorted in Increasing order or in Decreasing Order. While, when elements are popped from a simple queue, a FIFO order of data is obtained in the result.

Below is a simple implementation of the priority queue.

Python

# A simple implementation of Priority Queue
# using Queue.
class PriorityQueue(object):
    def __init__(self):
        self.queue = []
 
    def __str__(self):
        return ' '.join([str(i) for i in self.queue])
 
    # for checking if the queue is empty
    def isEmpty(self):
        return len(self.queue) == 0
 
    # for inserting an element in the queue
    def insert(self, data):
        self.queue.append(data)
 
    # for popping an element based on Priority
    def delete(self):
        try:
            max_val = 0
            for i in range(len(self.queue)):
                if self.queue[i] > self.queue[max_val]:
                    max_val = i
            item = self.queue[max_val]
            del self.queue[max_val]
            return item
        except IndexError:
            print()
            exit()
 
if __name__ == '__main__':
    myQueue = PriorityQueue()
    myQueue.insert(12)
    myQueue.insert(1)
    myQueue.insert(14)
    myQueue.insert(7)
    print(myQueue)           
    while not myQueue.isEmpty():
        print(myQueue.delete())

                    
Output:
12 1 14 7
14
12
7
1

Note that the time complexity of delete is O(n) in the above code. A better implementation is to use Binary Heap which is typically used to implement a priority queue. Note that Python provides heapq in the library also.

Time complexity: By using heap data structure to implement Priority Queues
Insert Operation: O(log(n))
Delete Operation: O(log(n))


Last Updated : 29 Aug, 2022
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads