Thread: UB-Tree

Hi,

The inventor of the B-Tree algorythm, R. Bayer, published 
in the past years ('96-'01) several scientific papers where 
he describes the new UB-Tree algorythm. It is based on 
ordinary B-Tree implementation but allows >multidimensional< 
indexing.

Apart from the need to update only one index instead of
several for each table insert/delete the UB-Tree indexing
allows to perform extremly performant lookups for queries 
that contain constraints for several columns (dimensions)
and sort criterias, like:

SELECT * FROM EMPLOYEES WHERE ID>10 AND ID<100 AND INCOME>5000 SORTED BY NAME

Such a query could profit from an UB-Tree index on the columns
(dimensions) ID, INCOME and NAME - all at once, using only 
one UB-Tree!

So the UB-Tree allows you:- to make range queries on all dimensions at once, and- read the result in sorted order (with
onlylittle overhead), where- each dimension can be read in sorted or reverse order
 

I guess such an indexing method would please any SQL database. 
Has anybody of you ever stumbled across the UB-Tree related
algorythms?

If not, you can download several PDF documents describing
the UB-Tree related algorythms from this URL (in english 
language!):

http://mistral.in.tum.de

I found no free implementation of the UB-Tree. The team
of R. Bayer only released closed source and sell it.

kind regards,
Robert Schrem

On Fri, 2002-03-08 at 20:48, Robert Schrem wrote:
> 
> If not, you can download several PDF documents describing
> the UB-Tree related algorythms from this URL (in english 
> language!):
> 
> http://mistral.in.tum.de

The last one tere seems to be the original explanation of the idea.
> I found no free implementation of the UB-Tree. The team
> of R. Bayer only released closed source and sell it.

This technique seems to be a good candidate for implementing using GiST
or perhaps just defined using the <. operator mentioned there. 

Mappign from ordinary query with several = , < and between may be a
little tricky though.

They may also have patents on it, so we should move carefully here.

-------------
Hannu

----- Original Message -----
From: "Robert Schrem" <robert@schrem.de>
To: "Hannu Krosing" <hannu@krosing.net>
Cc: "PostgreSQL Hackers List" <pgsql-hackers@postgresql.org>
Sent: Monday, March 11, 2002 11:55 AM
Subject: Fwd: Re: [HACKERS] UB-Tree
> On Friday 08 March 2002 20:37, Hannu Krosing  wrote:
> > They may also have patents on it, so we should move carefully here.
>
> I sent a mail asking R. Bayer about any known patent issues.
> He said that the UB-Tree is internationally patented.
> Sad, because it looked like a briliant idea. Now it looks like
> it will be banned from the open source community for some
> decades to come... :-(

IANAL, but there seem to be some issues :

1. There is no such thing as 'internationally patented' as most countries
still don't allow patenting software algorithms.

2. I doubt it is possible to patent a general idea, like replacing multiple
one-dimensional indexes with one multi-dimensional index (conceptually
they _are_ the same thing, just optimised for different access paterns)

So as we can't use UB-Tree, we may well achieve similar result buy
using a multi-dimensional/multi-type R-tree and teach our optimiser
to use it for resolving multiple where clause restrictions simultaneously.

Of course this too may be patented. If it is not, let's hope this e-mail
will be archived and can be used as prior art to prevent future patenting
attempts :)

Another common way of fighting such patent's is patenting all possible
future improvements and then haggle with the patent holder .

I think this could be something that is effectively doable by open-source
community, at least the part of generating the ideas.

-------------
Hannu