Jump to content
Larry Ullman's Book Forums

Recommended Posts

Hello Forum!

 

Asking yet another one of my typical questions.

 

Do I understand it correctly that the underlying notion of the "normal form" databases is to eliminate repetition from every database, so as a result there's a bigger number of smaller databases?

 

I'm asking this question because in Chapter 3 of the 3rd edition of this book Larry is using a database that contains repeated values in the same column:

 

dbexample.jpg

 

 

I also often find myself using databases like that, which seems to work perfectly fine.

 

Am I missing something, as always? Do I understand the idea of the "normal form" database incorrectly, or should there be (if we deal with the example above) a separate databases for all UNIQUE, STANDARD and PO BOX zip code types, and a separate databases for HOTSVILLE, AGUADILLA, ARECIBO and so on?

 

Would appreciate your input!

Link to comment
Share on other sites

Yes, your understanding is correct, although you're using the word "database" when you should be using "table".

 

As for that example table, it comes from a third party, so it's not my design. To normalize it, one would create separate tables for the type, city, and state. That being said, I don't always normalize to the nth degree and I would create a table like that in the real world myself.

  • Upvote 1
Link to comment
Share on other sites

Remember that a String can also be used as a key. "NY" can be used a primary key in the city table, and a foreign key in the table above. (General clarification not tied to normalization)

 

I would suggest you to learn the normalization rules, as it will help you to understand when you can break them.

  • Upvote 1
Link to comment
Share on other sites

 Share

×
×
  • Create New...