Paul Rubin wrote: >> items = d.items() >> items.sort(key=operator.itemgetter(1), reverse=True) >> >> the items list would require a couple of megabytes for 150k dictionary >> entries, or so. the key map needs some memory too, but the rest of >> the sort is done in place. > > I think the OP's method avoided the key map.
right. I should have written "most efficient", not memory efficient. the items+itemgetter approach is a bit faster, but uses a few extra megabytes of memory temporarily, during the sort. </F> -- http://mail.python.org/mailman/listinfo/python-list
