Re: Encoding materialized path in an atomic value.

"David Cressey" <david.cressey_at_earthlink.net> wrote in message news:ewZYe.2909$vw6.917_at_newsread1.news.atl.earthlink.net... .
..
> This only works for binary trees, but the same technique can be
> generalized, or a technique can be used for reducing arbitrary trees to
> binary tree equivalents.

There is one-to-one mapping between binary trees and ordered trees. So you can convert an arbitrary tree to an ordered tree by imposing some kind of [artificial] order amongst siblings, and then to a binary one at the expense of changing the original tree topology -- the first sibling becomes the new parent for the second sibling, and so on.

By the way, representing the materialized path as '1.22.7.16' rather than 'john.kate.bob.sam' is an example of converting an unordered tree to an ordered one.

>
> 23 is, I think we will all agree "atomic" or "simple" enough so that it
> can be stored in a single value. This case is independent of whether or
> not
> one agrees that only atomic values can be stored in a relational table.
>
>
>
>
>
>
>
>
>
Received on Sat Sep 24 2005 - 15:41:31 CEST