Steven D'Aprano <steve+comp.lang.pyt...@pearwood.info>: > If you are in a position to randomize the data before storing it in > the tree, an unbalanced binary tree is a solid contender.
Applications that can assume randomly distributed data are exceedingly rare and even then, you can't easily discount the possibility of worst-case ordering. In fact, Dan himself said unbalanced trees performed so badly he couldn't test them satisfactorily. Marko -- https://mail.python.org/mailman/listinfo/python-list