In the Python docs,

The lowest valued entries are retrieved first (the lowest valued entry is the one returned by sorted(list(entries))[0]). A typical pattern for entries is a tuple in the form: (priority_number, data).

It appears the queue will be sorted by priority then data, which may not be always correct. Suppose data “item 2”, is enqueued before “item 1”, item 1 will still go first. In another docs page, heapq, it suggests the use of a counter. So I will store my data like entry = [priority, count, task]. Isn’t there something like

PriorityQueue.put(item, priority)

Then I won’t need to implement the ordering myself?

Just use the second item of the tuple as a secondary priority if a alphanumeric sort on your string data isn’t appropriate. A date/time priority would give you a priority queue that falls back to a FIFIO queue when you have multiple items with the same priority. Here’s some example code with just a secondary numeric priority. Using a datetime value in the second position is a pretty trivial change, but feel free to poke me in comments if you’re not able to get it working.


import Queue as queue

prio_queue = queue.PriorityQueue()
prio_queue.put((2, 8, 'super blah'))
prio_queue.put((1, 4, 'Some thing'))
prio_queue.put((1, 3, 'This thing would come after Some Thing if we sorted by this text entry'))
prio_queue.put((5, 1, 'blah'))

while not prio_queue.empty():
    item = prio_queue.get()
    print('%s.%s - %s' % item)


1.3 - This thing would come after Some Thing if we didn't add a secondary priority
1.4 - Some thing
2.8 - super blah
5.1 - blah


Here’s what it looks like if you use a timestamp to fake FIFO as a secondary priority using a date. I say fake because it’s only approximately FIFO as entries that are added very close in time to one another may not come out exactly FIFO. I added a short sleep so this simple example works out in a reasonable way. Hopefully this helps as another example of how you might get the ordering you’re after.

import Queue as queue
import time

prio_queue = queue.PriorityQueue()
prio_queue.put((2, time.time(), 'super blah'))
prio_queue.put((1, time.time(), 'This thing would come after Some Thing if we sorted by this text entry'))
prio_queue.put((1, time.time(), 'Some thing'))
prio_queue.put((5, time.time(), 'blah'))

while not prio_queue.empty():
    item = prio_queue.get()
    print('%s.%s - %s' % item)

As far as I know, what you’re looking for isn’t available out of the box. Anyway, note that it wouldn’t be hard to implement:

from Queue import PriorityQueue

class MyPriorityQueue(PriorityQueue):
    def __init__(self):
        self.counter = 0

    def put(self, item, priority):
        PriorityQueue.put(self, (priority, self.counter, item))
        self.counter += 1

    def get(self, *args, **kwargs):
        _, _, item = PriorityQueue.get(self, *args, **kwargs)
        return item

queue = MyPriorityQueue()
queue.put('item2', 1)
queue.put('item1', 1)

print queue.get()
print queue.get()

Example output:


I made something like this to accomplish a FIFO similar to gfortune, but without the need of calling time.time() everywhere: (Python 3 only)

import time
from dataclasses import dataclass, field

class PrioritizedItem:
    prio: int
    timestamp: float = field(init=False, default_factory=time.time)
    data: object = field(compare=False)

Now you can do:

import queue

item1 = PrioritizedItem(0, "hello world")
item2 = PrioritizedItem(0, "what ever")
q = queue.PriorityQueue()

And be sure, they will always be extracted in the same order.