Lecture 17 Part 5 BCNF and other Normal Forms

[@gecicihesap901]

3:47 SHOTS FIRED

[@madinakamolova9283]

My textbook mentions that all types of normalization consider candidate keys (keys that are potential primary keys, but which were not chosen to be the PK or part of it) in any table. Here, if a table has many candidate keys which can determine other attributes (that are not candidate keys, a.k.a. "non-prime" attributes), then the BCNF is violated. So, naturally, in BCNF each table must have only one candidate key. Now, the PK is also a candidate key. Therefore, each table in BCNF must have only one candidate key, which PK is--by default. Why did I describe the logic from my textbook that may seem confusing and not entirely related to the video? Because it seems that BCNF is the same thing as 3NF, only without any candidate keys aside from the PK. If that is the case, then the explanation of BCNF has been made unnecessarily more complicated than it needs to be. Thank you for reading :)