PyMOTW: threading

The threading module lets you run multiple operations concurrently in the same process space.

Module: threading
Purpose: Builds on the thread module to more easily manage several threads of execution.
Python Version: since 1.5.2 (some of these examples require 2.5 because they use the with statement)

Description:

The threading module builds on the low-level features of the thread module to make working with threads even easier and more pythonic.

Thread objects:

The simplest way to use a thread is to instantiate it with a target function and call start() to let it begin working.

import threading

def worker():
    """thread worker function"""
    print 'Worker'
    return

threads = []
for i in range(5):
    t = threading.Thread(target=worker)
    threads.append(t)
    t.start()

The output, is unsurprisingly, 5 lines with "Worker" on each:

$ python threading_simple.py
Worker
Worker
Worker
Worker
Worker

It useful to be able to spawn a thread and pass it arguments to tell it what work to do. For example, in PyMOTW: Queue, I created a simple program to illustrate how to download enclosures from RSS/Atom feeds. Each downloader thread needed to know where to find the URLs, and the Queue instance was passed as an argument when the thread was created. Here, we'll just pass the thread a number so the output is a little more interesting in the second example.

import threading

def worker(num):
    """thread worker function"""
    print 'Worker:', num
    return

threads = []
for i in range(5):
    t = threading.Thread(target=worker, args=(i,))
    threads.append(t)
    t.start()

The integer argument is now included in the message printed by each thread:

$ python threading_simpleargs.py
Worker: 0
Worker: 1
Worker: 2
Worker: 3
Worker: 4

Determining the current thread:

Using arguments to identify or name the thread is cumbersome, and unnecessary. Each Thread instance has a name with a default value that you can change as the thread is created. Naming threads is useful if you have a server process with multiple service threads handling different operations.

import threading
import time

def worker():
    print threading.currentThread().getName(), 'Starting'
    time.sleep(2)
    print threading.currentThread().getName(), 'Exiting'

def my_service():
    print threading.currentThread().getName(), 'Starting'
    time.sleep(3)
    print threading.currentThread().getName(), 'Exiting'

t = threading.Thread(name='my_service', target=my_service)
w = threading.Thread(name='worker', target=worker)
w2 = threading.Thread(target=worker) # use default name

w.start()
w2.start()
t.start()

The debug output includes the name of the current thread on each line. The lines with "Thread-1" in the thread name column correspond to the unnamed thread w2.

$ python threading_names.py
worker Starting
Thread-1 Starting
my_service Starting
worker Exiting
Thread-1 Exiting
my_service Exiting

Of course, in most programs you won't use print to debug. The logging module supports embedding the thread name in every log message using the formatter code %(threadName)s. Including thread names in log messages makes it easier to trace those messages back to their source.

import logging
import threading
import time

logging.basicConfig(level=logging.DEBUG,
                    format='[%(levelname)s] (%(threadName)-10s) %(message)s',
                    )

def worker():
    logging.debug('Starting')
    time.sleep(2)
    logging.debug('Exiting')

def my_service():
    logging.debug('Starting')
    time.sleep(3)
    logging.debug('Exiting')

t = threading.Thread(name='my_service', target=my_service)
w = threading.Thread(name='worker', target=worker)
w2 = threading.Thread(target=worker) # use default name

w.start()
w2.start()
t.start()

The output from the format string above looks like:

$ python threading_names_log.py
[DEBUG] (worker    ) Starting
[DEBUG] (Thread-1  ) Starting
[DEBUG] (my_service) Starting
[DEBUG] (worker    ) Exiting
[DEBUG] (Thread-1  ) Exiting
[DEBUG] (my_service) Exiting

Daemon vs. Non-Daemon Threads:

Up until this point, I have been assuming that the main program does not exit until all threads have completed their work. Sometimes you will want to spawn a thread as a "daemon" that runs without blocking the main program from exiting. Using daemon threads is useful for services where there may not be an easy way to interrupt the thread or where letting the thread die in the middle of its work does not lose or corrupt data (for example, a thread that generates "heart beats" for a service monitoring tool). To mark a thread as a daemon, call its setDaemon() with a boolean argument. The default is for threads to not be daemons, so passing True turns the daemon mode on.

import threading
import time

def daemon():
    print 'Starting:', threading.currentThread().getName()
    time.sleep(2)
    print 'Exiting :', threading.currentThread().getName()

d = threading.Thread(name='daemon', target=daemon)
d.setDaemon(True)

def non_daemon():
    print 'Starting:', threading.currentThread().getName()
    print 'Exiting :', threading.currentThread().getName()

t = threading.Thread(name='non-daemon', target=non_daemon)

d.start()
t.start()

Notice that the output does not include the "Exiting" message from the daemon thread, since all of the non-daemon threads (including the main thread) exit before the daemon thread wakes up from its 2 second sleep.

$ python threading_daemon.py
Starting: daemon
Starting: non-daemon
Exiting : non-daemon

To wait until the daemon thread has completed its work, use the join() method.

import threading
import time

def daemon():
    print 'Starting:', threading.currentThread().getName()
    time.sleep(2)
    print 'Exiting :', threading.currentThread().getName()

d = threading.Thread(name='daemon', target=daemon)
d.setDaemon(True)

def non_daemon():
    print 'Starting:', threading.currentThread().getName()
    print 'Exiting :', threading.currentThread().getName()

t = threading.Thread(name='non-daemon', target=non_daemon)

d.start()
t.start()

d.join()
t.join()

Since we wait for the daemon thread to exit using join(), we do see its "Exiting" message.

$ python threading_daemon_join.py
Starting: daemon
Starting: non-daemon
Exiting : non-daemon
Exiting : daemon

By default, join() blocks indefinitely. It is also possible to pass a timeout argument (a float representing the number of seconds to wait for the thread to become inactive). If the thread does not complete within the timeout period, join() returns anyway.

import threading
import time

def daemon():
    print 'Starting:', threading.currentThread().getName()
    time.sleep(2)
    print 'Exiting :', threading.currentThread().getName()

d = threading.Thread(name='daemon', target=daemon)
d.setDaemon(True)

def non_daemon():
    print 'Starting:', threading.currentThread().getName()
    print 'Exiting :', threading.currentThread().getName()

t = threading.Thread(name='non-daemon', target=non_daemon)

d.start()
t.start()

d.join(1)
print 'd.isAlive()', d.isAlive()
t.join()

Since the timeout passed is less than the amount of time the daemon thread sleeps, the thread is still "alive" after join() returns.

$ python threading_daemon_join_timeout.py
Starting: daemon
Starting: non-daemon
Exiting : non-daemon
d.isAlive() True

Using enumerate() to wait for all running threads:

It is not necessary to retain an explicit handle to all of the daemon threads you start in order to ensure they have completed before exiting the main process. threading.enumerate() returns a list of active Thread instances. The list includes the current thread, and since joining the current thread is not allowed (it introduces a deadlock situation), we must check before joining.

import random
import threading
import time

def worker():
    """thread worker function"""
    t = threading.currentThread()
    pause = random.randint(1,5)
    print 'Starting:', t.getName(), 'sleeping', pause
    time.sleep(pause)
    print 'Ending  :', t.getName()
    return

for i in range(3):
    t = threading.Thread(target=worker)
    t.setDaemon(True)
    t.start()

main_thread = threading.currentThread()
for t in threading.enumerate():
    if t is main_thread:
        continue
    print 'Joining :', t.getName()
    t.join()

Since the worker is sleeping for a random amount of time, your output may vary. It should look something like this, though:

$ python threading_enumerate.py
Starting: Thread-1 sleeping 2
Starting: Thread-2 sleeping 5
Starting: Thread-3 sleeping 2
Joining : Thread-1
Ending  : Thread-1
Joining : Thread-3
Ending  : Thread-3
Joining : Thread-2
Ending  : Thread-2

Creating your own Thread class:

When you start a Thread, it does some basic setup and then calls its run() method, which calls the target function passed to the constructor. If you want to create your own type of thread, you can subclass from Thread and override run() to do whatever you want.

import threading

class MyThread(threading.Thread):

    def run(self):
        print 'MyThread:', self.getName()
        return

for i in range(5):
    t = MyThread()
    t.start()

$ python threading_subclass.py
MyThread: Thread-1
MyThread: Thread-2
MyThread: Thread-3
MyThread: Thread-4
MyThread: Thread-5

Starting a task in a thread with a Timer:

One example of a reason to subclass Thread is provided by Timer, also included in threading. A Timer lets you start the work of your thread after a delay, and cancel the operation at any point within that time period.

import threading
import time

def delayed():
    print 'Worker running', threading.currentThread().getName()
    return

t1 = threading.Timer(3, delayed)
t1.setName('t1')
t2 = threading.Timer(3, delayed)
t2.setName('t2')

print 'Starting timers'
t1.start()
t2.start()

print 'Waiting before canceling', t2.getName()
time.sleep(2)
print 'Canceling', t2.getName()
t2.cancel()
print 'Main thread done'

Notice that the second timer is never run, and the first timer appears to run after the rest of the main program is done. Since it is not a daemon thread, we do not have to join() it explicitly to block waiting for it.

$ python threading_timer.py
Starting timers
Waiting before canceling t2
Canceling t2
Main thread done
Worker running t1

Signaling between threads with Event objects:

Although the point of using multiple threads is to spin separate operations off to run more or less simultaneously, there are times when it is important to be able to synchronize the operations in two or more threads. A simple way to communicate between threads is using Event objects. An Event manages an internal flag that users can either set() or clear(). Other users can wait() for the flag to be set(), effectively blocking progress until allowed to continue. You can also think of an Event as a traffic light.

import logging
import threading
import time

logging.basicConfig(level=logging.DEBUG,
                    format='%(asctime)s (%(threadName)-10s) %(message)s',
                    )
                    
def wait_for_event(e):
    """Wait for the event to be set before doing anything"""
    logging.debug('wait_for_event starting')
    e.wait()
    logging.debug('e.isSet()->%s', e.isSet())

def wait_for_event_timeout(e, t):
    """Wait t seconds and then timeout"""
    logging.debug('wait_for_event_timeout starting')
    e.wait(t)
    logging.debug('e.isSet()->%s', e.isSet())


e = threading.Event()
t1 = threading.Thread(name='block', 
                      target=wait_for_event,
                      args=(e,))
t1.start()

t2 = threading.Thread(name='non-block', 
                      target=wait_for_event_timeout, 
                      args=(e, 2))
t2.start()

logging.debug('Waiting before calling Event.set()')
time.sleep(3)
e.set()
logging.debug('Event is set')

In this case, the non-block thread times out before the Event is set.

$ python threading_event.py
2008-01-13 13:25:02,514 (block     ) wait_for_event starting
2008-01-13 13:25:02,525 (non-block ) wait_for_event_timeout starting
2008-01-13 13:25:02,536 (MainThread) Waiting before calling Event.set()
2008-01-13 13:25:04,526 (non-block ) e.isSet()->False
2008-01-13 13:25:05,563 (MainThread) Event is set
2008-01-13 13:25:05,564 (block     ) e.isSet()->True

Controlling access to resources with Lock:

In addition to synchronizing the operations of threads, it is also important to be able to control access to shared resources to prevent corruption or missed data. Python's built-in data structures (lists, dictionaries, etc.) are thread-safe as a side-effect of having atomic byte-codes for manipulating them (so the GIL is not released in the middle of an update). Your own data structures implemented in Python (or simpler types like integers and floats), don't have that protection. To guard against simultaneous access to an object, use a Lock object.

import logging
import random
import threading
import time

logging.basicConfig(level=logging.DEBUG,
                    format='(%(threadName)-10s) %(message)s',
                    )
                    
class Counter(object):
    def __init__(self, start=0):
        self.lock = threading.Lock()
        self.value = start
    def increment(self):
        self.lock.acquire()
        try:
            logging.debug('Acquired lock')
            self.value = self.value + 1
        finally:
            self.lock.release()

def worker(c):
    for i in range(3):
        pause = random.random()
        logging.debug('Sleeping %0.02f', pause)
        time.sleep(pause)
        c.increment()
    logging.debug('Done')

counter = Counter()
for i in range(5):
    t = threading.Thread(target=worker, args=(counter,))
    t.start()

logging.debug('Waiting for worker threads')
main_thread = threading.currentThread()
for t in threading.enumerate():
    if t is not main_thread:
        t.join()
logging.debug('Counter: %d', counter.value)

In this example, the worker() function increments a Counter() instance. The Counter manages a Lock to prevent two threads from changing its internal state at the same time. If the Lock was not used, there is a possibility of missing a change to the value attribute.

The random module is used to introduce variable sleep durations for each time through the loop, so the output you see when running the sample code may vary.

$ python threading_lock.py
(Thread-1  ) Sleeping 0.66
(Thread-2  ) Sleeping 0.83
(Thread-3  ) Sleeping 0.41
(Thread-4  ) Sleeping 0.32
(Thread-5  ) Sleeping 0.77
(MainThread) Waiting for worker threads
(Thread-4  ) Acquired lock
(Thread-4  ) Sleeping 0.54
(Thread-3  ) Acquired lock
(Thread-3  ) Sleeping 0.18
(Thread-3  ) Acquired lock
(Thread-3  ) Sleeping 0.02
(Thread-3  ) Acquired lock
(Thread-3  ) Done
(Thread-1  ) Acquired lock
(Thread-1  ) Sleeping 0.22
(Thread-5  ) Acquired lock
(Thread-5  ) Sleeping 0.34
(Thread-2  ) Acquired lock
(Thread-2  ) Sleeping 0.51
(Thread-1  ) Acquired lock
(Thread-1  ) Sleeping 0.03
(Thread-4  ) Acquired lock
(Thread-4  ) Sleeping 0.26
(Thread-1  ) Acquired lock
(Thread-1  ) Done
(Thread-4  ) Acquired lock
(Thread-4  ) Done
(Thread-5  ) Acquired lock
(Thread-5  ) Sleeping 0.02
(Thread-5  ) Acquired lock
(Thread-5  ) Done
(Thread-2  ) Acquired lock
(Thread-2  ) Sleeping 0.17
(Thread-2  ) Acquired lock
(Thread-2  ) Done
(MainThread) Counter: 15

If another thread has acquired the lock, it might be preferable to find that out without blocking all action in the current thread. In that case, pass a False value for the blocking argument to acquire(). In the next example, worker() tries to acquire the lock 3 separate times, and counts how many attempts it has to make to do so. The random sleeps are used to simulate varying load in the different threads.

import logging
import random
import threading
import time

logging.basicConfig(level=logging.DEBUG,
                    format='%(asctime)s (%(threadName)-8s) %(message)s',
                    )
                    
def random_pause():
    pause = random.random()
    time.sleep(pause)                
                    
def worker(lock):
    logging.debug('Starting')
    num_tries = 0
    num_acquires = 0
    while num_acquires < 3:
        random_pause()
        have_it = lock.acquire(0)
        try:
            num_tries += 1
            if have_it:
                logging.debug('Iteration %d: Acquired',  num_tries)
                num_acquires += 1
                random_pause()
            else:
                logging.debug('Iteration %d: Not acquired', num_tries)
        finally:
            if have_it:
                lock.release()
    logging.debug('Done after %d iterations', num_tries)

lock = threading.Lock()
for i in range(3):
    t = threading.Thread(target=worker, args=(lock,))
    t.start()

As you can see, it takes some of the threads many more than 3 iterations to acquire the lock 3 separate times.

$ python threading_lock_noblock.py
2008-01-13 14:48:27,712 (Thread-1) Starting
2008-01-13 14:48:27,723 (Thread-2) Starting
2008-01-13 14:48:27,735 (Thread-3) Starting
2008-01-13 14:48:28,173 (Thread-1) Iteration 1: Acquired
2008-01-13 14:48:28,624 (Thread-2) Iteration 1: Acquired
2008-01-13 14:48:28,674 (Thread-3) Iteration 1: Not acquired
2008-01-13 14:48:28,861 (Thread-1) Iteration 2: Not acquired
2008-01-13 14:48:29,314 (Thread-2) Iteration 2: Acquired
2008-01-13 14:48:29,653 (Thread-3) Iteration 2: Not acquired
2008-01-13 14:48:29,674 (Thread-1) Iteration 3: Not acquired
2008-01-13 14:48:29,986 (Thread-1) Iteration 4: Not acquired
2008-01-13 14:48:30,003 (Thread-1) Iteration 5: Not acquired
2008-01-13 14:48:30,067 (Thread-3) Iteration 3: Not acquired
2008-01-13 14:48:30,543 (Thread-2) Iteration 3: Acquired
2008-01-13 14:48:30,605 (Thread-2) Done after 3 iterations
2008-01-13 14:48:30,639 (Thread-3) Iteration 4: Acquired
2008-01-13 14:48:30,755 (Thread-1) Iteration 6: Not acquired
2008-01-13 14:48:31,069 (Thread-1) Iteration 7: Acquired
2008-01-13 14:48:31,943 (Thread-3) Iteration 5: Not acquired
2008-01-13 14:48:32,654 (Thread-3) Iteration 6: Acquired
2008-01-13 14:48:32,949 (Thread-1) Iteration 8: Not acquired
2008-01-13 14:48:33,693 (Thread-1) Iteration 9: Acquired
2008-01-13 14:48:33,812 (Thread-1) Done after 9 iterations
2008-01-13 14:48:34,336 (Thread-3) Iteration 7: Acquired
2008-01-13 14:48:35,003 (Thread-3) Done after 7 iterations

Re-entrant Locks:

Normal Lock objects cannot be acquired more than once, even by the same thread. This can introduce undesirable side-effects if a lock is accessed by more than one function in the same call chain.

import threading

lock = threading.Lock()

def first():
    print 'First try:', lock.acquire()
    second()
    
def second():
    print 'Second try:', lock.acquire(0)

first()

In this case, since both functions are using the same global lock, and one calls the other, the second acquisition fails and would have blocked if we did not tell acquire() not to block.

$ python threading_lock_reacquire.py
First try: True
Second try: False

In a situation where separate code from the same thread needs to "re-acquire" the lock, use an RLock instead.

import threading

lock = threading.RLock()

def first():
    print 'First try:', lock.acquire()
    second()
    
def second():
    print 'Second try:', lock.acquire(0)

first()

The only change to the code from the previous example was substituting RLock for Lock.

$ python threading_rlock.py
First try: True
Second try: 1

Locks and with:

Locks are also compatible with new with statement, introduced in __future__ for Python 2.5 and part of the language for 2.6. Using with removes the need to explicitly acquire and release the lock.

from __future__ import with_statement
import threading

def worker_with(lock):
    with lock:
        print 'Lock acquired via with'
        
def worker_no_with(lock):
    lock.acquire()
    try:
        print 'Lock acquired directly'
    finally:
        lock.release()

lock = threading.Lock()
w = threading.Thread(target=worker_with, args=(lock,))
nw = threading.Thread(target=worker_no_with, args=(lock,))

w.start()
nw.start()

The two functions worker_with() and worker_no_with() manage the lock in equivalent ways.

$ python threading_lock_with.py
Lock acquired via with
Lock acquired directly

Synchronizing threads with a Condition object:

In addition to using Events, another way of synchronizing threads is through using a Condition object. Because the Condition uses a Lock, it can be tied to a shared resource. This allows threads to wait for the resource to be updated. In this example, the consumer() threads wait() for the Condition to be set before continuing. The producer() thread is responsible for setting the condition and notifying the other threads once they can continue.

from __future__ import with_statement
import logging
import threading
import time

logging.basicConfig(level=logging.DEBUG,
                    format='%(asctime)s (%(threadName)-2s) %(message)s',
                    )

def consumer(cond):
    """wait for the condition and use the resource"""
    logging.debug('Starting consumer thread')
    t = threading.currentThread()
    with cond:
        cond.wait()
        logging.debug('Resource is available to consumer')

def producer(cond):
    """set up the resource to be used by the consumer"""
    logging.debug('Starting producer thread')
    with cond:
        logging.debug('Making resource available')
        cond.notifyAll()

condition = threading.Condition()
c1 = threading.Thread(name='c1', target=consumer, args=(condition,))
c2 = threading.Thread(name='c2', target=consumer, args=(condition,))
p = threading.Thread(name='p', target=producer, args=(condition,))

c1.start()
time.sleep(2)
c2.start()
time.sleep(2)
p.start()

The threads use with to acquire the lock associated with the Condition. You can also call the acquire() and release() methods explicitly, if you are not using Python 2.5.

$ python threading_condition.py
2008-01-13 13:01:24,226 (c1) Starting consumer thread
2008-01-13 13:01:26,238 (c2) Starting consumer thread
2008-01-13 13:01:28,248 (p ) Starting producer thread
2008-01-13 13:01:28,249 (p ) Making resource available
2008-01-13 13:01:28,249 (c1) Resource is available to consumer
2008-01-13 13:01:28,250 (c2) Resource is available to consumer

Controlling concurrent access to resources with a Semaphore:

Sometimes it is useful to allow more than one access to a resource at a time, while still limiting the overall number. For example, a connection pool might support a fixed number of simultaneous connections, or a network application might support a fixed number of concurrent downloads. A Semaphore is one way to manage those connections.

from __future__ import with_statement
import logging
import random
import threading
import time

logging.basicConfig(level=logging.DEBUG,
                    format='%(asctime)s (%(threadName)-2s) %(message)s',
                    )

class ActivePool(object):
    def __init__(self):
        super(ActivePool, self).__init__()
        self.active = []
        self.lock = threading.Lock()
    def makeActive(self, name):
        with self.lock:
            self.active.append(name)
    def makeInactive(self, name):
        with self.lock:
            self.active.remove(name)
    def __str__(self):
        with self.lock:
            return str(self.active)

def worker(s, pool):
    with s:
        name = threading.currentThread().getName()
        pool.makeActive(name)
        logging.debug('Running: %s', str(pool))
        time.sleep(random.random())
        pool.makeInactive(name)

pool = ActivePool()
s = threading.Semaphore(5)
for i in range(20):
    t = threading.Thread(target=worker, name=str(i), args=(s, pool))
    t.start()

In this example, the ActivePool class simply serves as a convenient way to track which threads are able to run at a given moment. A real resource pool would probably allocate a connection or some other value to the newly active thread, and reclaim the value when the thread is done. Here it is just used to hold the names of the active threads to show that only 5 are running concurrently.

$ python threading_semaphore.py
2008-01-13 15:35:06,562 (0 ) Running: ['0']
2008-01-13 15:35:06,573 (1 ) Running: ['0', '1']
2008-01-13 15:35:06,584 (2 ) Running: ['0', '1', '2']
2008-01-13 15:35:06,596 (3 ) Running: ['1', '2', '3']
2008-01-13 15:35:06,607 (4 ) Running: ['1', '2', '3', '4']
2008-01-13 15:35:06,619 (5 ) Running: ['1', '2', '3', '4', '5']
2008-01-13 15:35:06,671 (6 ) Running: ['1', '3', '4', '5', '6']
2008-01-13 15:35:06,729 (7 ) Running: ['1', '3', '4', '5', '7']
2008-01-13 15:35:07,139 (8 ) Running: ['1', '3', '4', '7', '8']
2008-01-13 15:35:07,502 (9 ) Running: ['1', '3', '4', '7', '9']
2008-01-13 15:35:07,564 (10) Running: ['3', '4', '7', '9', '10']
2008-01-13 15:35:07,566 (11) Running: ['9', '10', '11']
2008-01-13 15:35:07,566 (12) Running: ['9', '10', '11', '12']
2008-01-13 15:35:07,567 (13) Running: ['9', '10', '11', '12', '13']
2008-01-13 15:35:07,683 (14) Running: ['9', '11', '12', '13', '14']
2008-01-13 15:35:08,055 (15) Running: ['11', '12', '13', '14', '15']
2008-01-13 15:35:08,149 (16) Running: ['12', '13', '14', '15', '16']
2008-01-13 15:35:08,150 (17) Running: ['12', '13', '15', '16', '17']
2008-01-13 15:35:08,319 (18) Running: ['12', '13', '16', '17', '18']
2008-01-13 15:35:08,531 (19) Running: ['12', '16', '17', '18', '19']

Keeping thread-specific data with local:

While some resources need to be locked so multiple threads can use them, others need to be protected so that they are hidden from view in threads that do not "own" them. The threading.local() creates an object capable of hiding values from view in separate threads.

import random
import threading

def show_value(data):
    print threading.currentThread().getName(), ': value=',
    try:
        print data.value
    except AttributeError:
        print 'No value yet'

def worker(data):
    show_value(data)
    data.value = random.randint(1, 100)
    show_value(data)

local_data = threading.local()
show_value(local_data)
local_data.value = 1000
show_value(local_data)

for i in range(2):
    t = threading.Thread(target=worker, args=(local_data,))
    t.start()

Notice that local_data.value is not present for any thread until it is set in that thread.

$ python threading_local.py
MainThread : value= No value yet
MainThread : value= 1000
Thread-1 : value= No value yet
Thread-1 : value= 19
Thread-2 : value= No value yet
Thread-2 : value= 26

To initialize the settings so all threads start with the same value, use a subclass and set the attributes in __init__().

import random
import threading

def show_value(data):
    print threading.currentThread().getName(), ': value=',
    try:
        print data.value
    except AttributeError:
        print 'No value yet'

def worker(data):
    show_value(data)
    data.value = random.randint(1, 100)
    show_value(data)

class MyLocal(threading.local):
    def __init__(self, value):
        print '(Initializing %s for %s)' % (id(self), threading.currentThread().getName()),
        self.value = value

local_data = MyLocal(1000)
show_value(local_data)

for i in range(2):
    t = threading.Thread(target=worker, args=(local_data,))
    t.start()

The output shows that __init__() is invoked on the same object (note the id() value), once in each thread.

$ python threading_local_defaults.py
(Initializing 479616 for MainThread) MainThread : value= 1000
Thread-1 : value= (Initializing 479616 for Thread-1) 1000
Thread-1 : value= 79
Thread-2 : value= (Initializing 479616 for Thread-2) 1000
Thread-2 : value= 56

References:

Python Module of the Week Home
Download Sample Code
The thread module
PyMOTW: Queue

Technorati Tags:
python, PyMOTW