Creative Commons baby!

Steven Danz wrote:

Wow... you mind if I borrow this? I'll reference you of course. I used to do the DBA thing long ago, and I've been trying to explain some of these same things to some DB newbs, and just couldn't get it to come across this clearly.

Thanks for your time.

Steven

Jim Herrmann wrote:

...

Tim, It sounds like you are wanting to do a fairly simple database structure, which is a good way to get started. When you say "open database format", I assume that you are talking about a target platform of a relational database, something like MySQL or Postgres. Yes? But that's getting ahead of the design curve. Here's how to proceed if you want to follow something akin to a formal database design process. I've been doing this for about 18 years as a DBA. I hope I'm not volunteering too much information. You said you were a newb. :-)

The first thing you need to do is create a logical model. When doing a logical model use English business names, and not programming abbreviations. Identify all the *Entities* that you need to represent, and the *Relationships* between those entities. A entity is a singular noun, which would include things like "Person", "Department", or in your case, "Reminder". A relationship is a verb phrase like "has many", "has one", or "has zero or many". So, you can say something like "a Department has many Persons" and "a Department has one and only one Manager".

Once you have this, you can create an Entity-Relationship (ER) diagram. This gives you a very high level view of your data structure. This would be a box for each entity and a line between the boxes represents a relationship. Exactly what the box and line looks like depends on the modeling tool you use. Then you need to identify all the attributes (fields) for each entity. You need to identify which attributes uniquely identify an occurrence (row) of that entity, and that's your natural key. Keys of parents, of a relationship, migrate to the children. If the parent is required for a child to exist, then it is an "identifying" relationship, and and the parent's keys become part of the child's keys. If it is a "non-identifying" relationship, then some or all of the parent key is NOT part of the child's key. Now, if you find an attribute occurs multiple times in an entity, then you should "normalize" those attributes into another entity or entities. There are several levels of normalization, and I won't get into it here as it would be a full day class, but third normal form is considered the best level of normalization for most DB designs. That's where each attribute of each entity is dependent upon the key, the whole key, and nothing but the key, so help me Codd. (E.F. Codd, an IBM fellow, created relational database theory.) :-)

Once you have a fully attributed ER diagram, you are ready to create a physical model based on the logical model you created. This is where you create names that you want to use in your program, as you turn entities into *tables* and attributes become *columns* and relationships become *foreign keys*. You might shorten "Department" to "DEPT", for example. It's best if you use standard abbreviations, meaning be consistent, as much as possible. It makes programming much easier. The length of table and column names may need to be shortened, depending on the target database. That's part of the physical design process. You also need to decide what indexes will be on your tables. Some for uniqueness, such as on the key, and some for performance, depending on how you are going to access your data.

There are also performance and convenience tricks that you can do when you take your model physical. For example, if you have some natural keys that have lots of columns, you might introduce artificial keys, which would be something like an integer or reverse timestamp. You may also at this point decide to "denormalise" some entities to make your system go faster, but that comes at a price of maintaining redundant data, and it breaks Codd's rule. Depending on the DB platform, there are lots of physical implementation issues to consider, if you are building something with lots of data or lots of transactions, but you probably don't need to concern yourself with all that for this project.

Hopefully you find this information helpful. If you have any further questions about data base, I'd be glad to help. Sorry about the brain dump. ;-)

Enjoy, Jim

Tim reid wrote:

...

I'm thinking of programing/writing/whatever a database to keep track of some information for work. I think that I would like to keep track of 7-12 fields of information, and be able to set up reminders that are tagged to date information. For instance, "you did such and such 30 days ago...you can now file the claim, here is the information that you recorded earlier."

How would i go about doing this, free as in beer, in an open database format? Let me also mention that I'm a total newb ;)

TIA,

Tim