Iterators vs. Generators

[Chris Liechti]

aahz at pythoncraft.com (Aahz) wrote in news:ae3a6f$9d9$1 at panix1.panix.com:

> So when would one actually write an iterator instead of a generator?
> I've been trying to think of an example and failing.

don't know, maybe when you don't like the additional loop in the generator 
(see example below)

maybe the iterator has advantages when the data in the object changes or 
has to be chaged as the object q (from below) _is_ the iteraror but it's 
_not_ the generator (but this does not matter in the example below ...).
if you pass "i = iter(q)" somewhere then with the iterator (and forget 
about q) you can still modify the original object as "i is q" (at least in 
my example), but this is not even possible with a generator.

iterators can be generated implicit by __iter__ in a "for", but a generator 
needs a factory function (of course that can be named __iter__ ...)

seems to be a matter of taste...

chris

PS: why is Queue not iterable by default?
here's a Queue for fun <wink>:
----

#!/usr/bin/env python
from __future__ import generators

import Queue, threading, time

class Q(Queue.Queue):
    def __iter__(self):
        return self
    def next(self):
        x = self.get(1)
        if x is None: raise StopIteration
        return x

class Q2(Queue.Queue):
    def __iter__(self):
        while 1:
            x = self.get(1)
            if x is None: return
            yield x
    

class Producer(threading.Thread):
    def __init__(self, queue):
        threading.Thread.__init__(self)
        self.q = queue
    def run(self):
        for n in range(10):
            time.sleep(0.2)
            self.q.put(n)
        self.q.put(None)

if __name__ == '__main__':
    #iterator
    q = Q()
    print ":", q is iter(q), type(iter(q))
    Producer(q).start()
    for x in q:
        print x
        
    #generator
    q = Q2()
    print ":", q is iter(q), type(iter(q))
    Producer(q).start()
    for x in q:
        print x

-- 
Chris <cliechti at gmx.net>