Here's a quick little hack that seems to work:

import boto import threading

class SimpleThread(threading.Thread):

def __init__(self, name, domain_name, item_names): threading.Thread.__init__(self, name=name) print 'SimpleThread: %s %s' % (name, item_names) self.domain_name = domain_name self.conn = boto.connect_sdb() self.item_names = item_names self.items = []

def run(self): for item_name in self.item_names: item = self.conn.get_attributes(self.domain_name, item_name) self.items.append(item) print 'got %s' % item.name

def get_attrs(domain_name, query, num_threads): item_names = [] conn = boto.connect_sdb() domain = conn.get_domain(domain_name) rs = domain.query(query) for item in rs: item_names.append(item.name) print item_names print len(item_names) items = [] threads = [] n = len(item_names) / num_threads for i in range(0, num_threads+1): thread = SimpleThread('Thread-%d' % i, domain_name, item_names[n*i:n*(i+1)]) threads.append(thread) thread.start() for thread in threads: thread.join() items = items + thread.items return items

So, you call the get_attr function with the domain_name, a query that you want to perform against that domain (could be empty string if you want all items in the domain) and the number of threads you want to start. It then collects all of the item names and fires up the required number of threads, passing each thread a list of item names it needs to retrieve. Each thread then creates it's own connection to SDB and gets the items associated with the item_names. Those are collected in a local list and when the threads are joined, each threads local list is concatenated to form a list of all items.

This code is very rough and will probably break in a number of interesting ways but for my simple tests it does work and, by golly, it's pretty darn fast!

On Jan 18, 3:53 pm, "Herb...@gmail.com" <Herb...@googlemail.com> wrote: