From: Chris R. <chr...@me...> - 2001-05-29 16:49:57
|
Graham Barr <gb...@po...> wrote: > It is taking so much memory because it is reading all the results into > memory before returning for you to process them. > > I suspect you don't really need all of the entries, so try modifying your > filter to result the resutl set that is being sent. > > If this is not possible, try using the callback option. This allows you > to process each entry as it arrives from the server. > > Graham. This is true, however 95000 * 200 bytes (rough size of the example entry) only comes to about 18 megabytes, and Simon said his perl process was using 250 megabytes :-( That's a lot of overhead, unless Simon missed a zero out of his entry count - 950000 entries would work out to use about the observed quantity of memory. Well, the same order of magnitude anyway. Cheers, Chris |