[Shacs] database structure question.. this one is easy

[Bronius Motekaitis]

Thanks all, for your answers thus far.  Be sure to do a ReplyAll or change
the Bronius to a SHACS so the world can bask in your intelligence.

So far, I have 2 for a defaults table, 1 opposed.  I'm leaning toward the
defaults table, too: in addition to making it clear that these are "default"
values without renaming foreign keys, it would also, should I so desire some
time in the future, the developer to allow multiple defaults per court.  Say
a given Court's Hearing is always held in one of two Locations 50% of the
time.  Putting defaults on the Court table would disallow for this expansion
without making a "dummy court entry" to reflect each of either case.  That
would be bad design..

To clarify one response: no, the 2nd example doesn't have the location,
judge, required.. in it, because those are just default/seed values.  The
whole point of the defaults table was that in creating a Hearing, pull a few
Hearing value defaults from somewhere-- these aren't defaults for the Court
table when creating a new Court.

* note: replace "Hearing" with "Trial" for my diagrams.. that was a recent
change in schema to better reflect the real world :)

About one fellow's comment on trial duration: yes, good point to consider.
However, for the scope of this project (it's an online
prequalification/impaneling system for jurors), trial duration can be 1 day
on through 7 or 14 days long.  1hr and 1/2 day sessions get rounded up.

-bronius


On Wed, Feb 25, 2009 at 8:58 AM, <James.D.Wyman at gmail.com> wrote:

> So you want to default values in a database? Why not create a trigger if
> these values are empty to insert a default value? I do not reccomend auto
> populating anything, instead have the input program generate the default and
> have a hardcoded default or a table you reference for default values for the
> program that are kept seperate from the data being stored. You should never
> have database driven programs that populate user data inside of themselves.
> I would say use two different sets of tables. One table for information to
> drive the program and pre-populate fields and your other database to store
> the information.
>
>
> On Feb 25, 2009 8:53am, Bronius Motekaitis <bronius.motekaitis at gmail.com>
> wrote:
> > Let's consider the tables: Hearing, Court, Judge, and Location (notice
> the CAPS: I did that so I wouldn't scare away any Access fans and
> undergrads..!)
> >
> > A Hearing can hear any Court's cases at any Location by any Judge, so
> these are all foreign keys on Hearing. When creating a Hearing, I want to
> populate default values for Judge, Location and a handful of other values
> not shown here based on the Court selected.
> >
> >
> > Where should I attach these default values?  Does it make more sense to
> make a "Court_Default" table or stuff the given Court table with these
> foreign key values as defaults?
> >
> > See attached for comparison of the two models considered..
> >
> >
> > -bronius
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.shsu.edu/pipermail/shacs/attachments/20090225/8eb8ec38/attachment.html