the past. Your best option would be #3
(produce a fully cross-computed table) because you can use indexed
lookups rather than computed queries... *much* faster, and worth the
massive amount of disk space it consumes.
--
Tom Bradford
The dbXML Project
Open Source Native XML Database
http://www
The table
and any associated indexes are completely dumped, and the process is
unrecoverable. In this one, the internal IDs will also remain the same.
--
Tom Bradford - http://www.tbradford.org
Developer - dbXML - http://www.dbxml.org
Maintainer - jEdit-Syntax - http://syntax.jedit.org
Co-
for any single database
operation. If you break the table out into multiple tables, segmented
into groups of entities that logically relate to each other, you can
actually achieve much better performance. Then, when you need some of
the data that's been isolated from the main table, you can