"Dj Gilcrease" <[EMAIL PROTECTED]> wrote >I was wondering if it was possible to find and object by it's ID.
Its probably possible but almost certainly the wrong thing to do... > what I want to do is something like > snipped... > The reason I want to do this is right now I store all the data in a > dict of class > > eg > class connections: > def __init__(self, *args, **kwargs): > self.sock = args[0] > self.myotherdata = stuff > ... > def incomingConnection(self, stuff): > (incSock, incAdder) = self.__sock.accept() > self.conns[incAdder] = connections(incSock, ...) > > > This takes up about 20 megs per person connecting, which is very > high But I'm not sure hpow storing IDs will help, the objects still need to be in memory. The only saving is in the actual dictionary itself which is relatively small. Also you could probably save memory more easily by storing the data to disk using pickle or shelve, but you'd need to be very sure that you could retrieve the objects fast enough to process your incoming connections - what is the peak rate? > IMHO, and I am hoping by storing all the data in the database and > only You only seemed to be storing the ID - which is basically the memory location. If you are going to store all the data in memory then the ID as a key is pointless you might as well get the database to generate one for you or use a simple counter at clkass level because you will need to regenerate the instance in memory to use it and that will result in it having a different ID... If you really need to reduce footprint you should probably consider connection object pooling where you use a cache of connection objects and reuse them. That way you only need the number of connections that are active at any one time - and you will always need those regardless of what you do with the inactive ones. Also look at the format of your data, is that in the most efficient format? Could it be encoded using struct for example? Or zipped? HTH, -- Alan Gauld Author of the Learn to Program web site http://www.freenetpages.co.uk/hp/alan.gauld _______________________________________________ Tutor maillist - Tutor@python.org http://mail.python.org/mailman/listinfo/tutor
