i want to unsubscribe. I have been email from sqlite and fill up my inbox
please unsubcribe this
From: Darren Duncan <[EMAIL PROTECTED]>
Reply-To: sqlite-users@sqlite.org
To: sqlite-users@sqlite.org
Subject: Re: [sqlite] Database design questions
Date: Tue, 27 Jun 2006 10:33:03 -0700
At 10:17 AM -0700 6/27/06, Bud Beacham wrote:
I am using SQLite to add a database engine to a product I hope
to sell. One thing that is desirable is to allow the customers
to enhance the database for their specific needs. For example,
they could add new tables to reflect their environment, and then
design queries that look at both the standard and added tables.
However, since this all new to me I have two questions.
If I offer an upgrade version with new tables how difficult will
it be for the user to merge their database into the latest
version? I will insure that any customer table names will not
conflict with my table names.
Would it make more sense to have the customers create a new
database and then design their queries to deal with two
databases?
Keeping everything that belongs together in a single disk file is usually
simpler for users. Within that constraint, the simplest way to incorporate
both your own tables plus arbitrary user-defined ones is for you to rename
the user-defined table names in a reversable fashion, to guarantee their
uniqueness from yours. The simplest way to do this is simply add a common
name prefix to all user-defined names behind the scenes, but hide that
implementation detail from the user; eg, prepend "udt_" (user defined
table) to each name the user picks, and make sure that none of your
application defined tables start with that. FYI, this practice simulates
multiple name-spaces / schemas in a single disk file, which SQLite doesn't
do natively. -- Darren Duncan
_________________________________________________________________
How good are you in a Formula One car? Play now
http://server1.msn.co.in/sp05/tataracing/onlinegame.asp
