Tony's tech terms: Relational database

We’ve mentioned Oracle databases a lot. Oracle is just one type of database called a relational database. A relational database is designed to limit the amount of duplicate data by splitting up different types of information by their types. The different types of data are stored in separate tables that have a defined relationship. That relationship makes the data easier to query.

For example:

• A campus will have many colleges
• A college will have many students
• A college’s students will have many courses

Now imagine this data in a spreadsheet. Each row would have a column for campus, college, student, and course.

Campus	College	Student	Course
UMNTC	CEHD	Example Student	Intro to Basket Weaving
UMNTC	CEHD	Example Student	Basket Weaving Lab

Those first three columns all have redundant data. This example doesn’t seem so bad, but if you multiply that by millions of students going back decades, that information starts to take up a lot of space. Not only that, but updating a single student might mean updating hundreds or thousands of rows. Splitting the data up into related tables like this is called normalization.

We can create a report like the spreadsheet by joining tables on their related IDs. It’s intuitive that the campus_id field in the colleges table maps to the ID field in the campus table. Joining them is fairly straightforward. We handle queries that are much more complicated than this simple example.

There are many other kinds of databases that structure their data differently depending on the use case. However, Oracle is the University-provided solution that holds all our Peoplesoft data. Retrieving that data is easier for us if we stay in that world. Besides, we don’t want to run and manage our own databases.