Keys are basically a unique way to identify a row in a table. They have a similar concept to a key that allows you to open a lock. Keys are unique and there are no copies. In fact, the database won't allow you to make a duplicate in the same table.

In relational database design, Keys help identify entities across different tables. For example, an Employee ID is a great way to uniquely identify an employee. The Key is what would "relate" across different tables, hence relational databases like SQL. You create relationships for your data in this way.

Keys are just normal fields with normal pieces of data. They can be numbers, letters, or a combination of the two.

You can create keyless tables and there are situations where that is okay. More than likely though, you'll want a unique key.

Relationships between your data happen when you have the same keys in the tables. Since they are the same keys, they relate to each, just like people with similar interests. If you tried to match people together that didn't have similar interests, they wouldn't relate. On the flip side, if you put people together that have the same interests, they will relate. It's the same with keys and relational data. You use the keys to create a relationship with data because they have the same interests.

For example, the same key data would exist in your employees table, your address table, and your phone number table. This would allow your queries to match rows in one table to another.