Xref: alice comp.databases.oracle.server:83008
Path: alice!news-feed.fnsi.net!newsfeed.icl.net!news-MUC.ecrc.net!blackbush.xlink.net!news-kar1.dfn.de!news-han1.dfn.de!newsserver.rrzn.uni-hannover.de!not-for-mail
From: Carsten Kleiner <ck@informatik.uni-hannover.de>
Newsgroups: comp.databases.oracle.server
Subject: Oracle 8i Spatial Problems
Date: Wed, 02 Feb 2000 10:09:22 +0100
Organization: RRZN - Newsserver Test
Lines: 26
Message-ID: <3897F442.59524314@informatik.uni-hannover.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: quoted-printable
X-Trace: newsserver.rrzn.uni-hannover.de 949482562 6863 130.75.26.4 (2 Feb 2000 09:09:22 GMT)
X-Complaints-To: abuse@newsserver.rrzn.uni-hannover.de
NNTP-Posting-Date: 2 Feb 2000 09:09:22 GMT
X-Mailer: Mozilla 4.71 [en] (X11; I; SunOS 5.7 sun4u)
X-Accept-Language: de, en

Hi,

we are experimenting a lot with Oracle 8i Spatial. We are experiencing
problems answering point-in-polygon queries with both SDO_RELATE and
SDO_GEOM.RELATE (object-relational spatial data model). The problem is
that for certain points (no matter if you store them as points or as
tiny rectangles) there is no polygon found that includes the point, but
the whole space is covered by polygons. =


Specifically there are certain polygons where one or more horizontal
lines of points within that object are reported to have empty
intersection. If we move those points (whose coordinates are 9 digit
integers) by 1 (virtually nothing) we obtain much fewer errors. It might
have something to do with the spatial index (maybe the points are
exactly on tile borders ?).

Has anyone experienced similar problems (maybe even solved those) ?
-- =

Carsten Kleiner

Institut f=FCr Informatik              Tel. +49-511-762-4960
FG Datenbanken + Informationssysteme Fax: +49-511-762-4961
Lange Laube 22                       Mail: ck@informatik.uni-hannover.de
D-30159 Hannover