Hi James,
Just remember what the idea of normalisation is: to make your data
consistent, prevent wasted space, and ease long term maintenance and
updates. So, make sure every table contains information about only one
'thing'. If columns in that table seem to repeat the same type of aspect of
that thing, then maybe theses aspects should be in another table. (e.g.
OfficeAddress, PostalAddres - should there be an address table?). Use
relationships to make sure that sub-things can't be orphaned from their
parent things. Index the fields used for joins so the normalised data can
be put back together again using queries as efficiently as possible.
If you've normalise the living daylights out of your data, then find that
you are continually recombining it, and this is affecting performance, then
consider de-normalising, but document what you've done and why so future
developers understand your decisions.
If you bear this kind of top level vision of database design in mind, then
all the complexities of normalisation will drop into context, and be a lot
easier to understand and use.
Good luck,
Andrew
"James W." <wa**@cwgsy.net> wrote in message
news:cb**************************@posting.google.c om...
Hi everyone,
I would like to ask you lot if you know of a great place on the
internet for a fool proof idiots guide to Normalisation 1st, 2nd, 3rd.
I looked at webopedia but that is just as confusing.
I am looking for a complete idiots guide that really is so simple all
it needs is pictures to turn it into a kiddies book
Thanks People
James W.