1 Class Agenda (11/07 and 11/12)  Review HW #8 answers  Present normalization process Enhance conceptual knowledge of database design. Improve practical skills of database design.  Approach Review what we learned from SQL. Review goals of database design. Identify and define vocabulary for normalization. Do an “intuitive” database design for a refresher. Discuss the characteristics of the three normal forms and the characteristics of a data model in third normal form. Use the normalization process to do the same database design. Compare results. Do more database design exercises using normalization process.

SQL and database design  Tables in a relational database aren’t really “related” unless the DBA wants to enforce referential integrity.  Tables are joined together by shared fields. SQL doesn’t care how you join tables together – it is completely up to the SQL programmer to make that decision.  It is really easy to get the wrong number of rows in the result table when joining tables.  It is easier to join tables correctly if you follow the relationship structure of a good database design. 2

3 Review of database design goals  Protect the integrity of the data. Reduce data redundancy. Prevent data anomalies.  Provide for change. Prevent inflexible data structures. Anticipate changes.  Provide access to complete data for decision making.

Data anomalies  An anomaly is a potential error or inconsistency in the data.  Data anomalies are most frequently caused by the implementation of M:N relationships.  M:N relationships are implemented through 1:M relationships that aren’t actually an accurate picture of the data... 4

Example of a “False” 1:m Relationship Order net net net cod cod Product 5613 tumbler 12 oz ea inventory Sacramento South tumbler 12 oz ea inventory Sacramento North food processor ea inventory Sacramento North paper rm supplies Sacramento South glass pitcher ea inventory Reno

6 Potential anomalies in the example  Insertion anomaly: Can’t add “some” of a row; must have all the key attributes. Example - Suppose we need to add a new warehouse?  Deletion anomaly: Lose some relevant data when deleting other data. Example - What happens to the Reno warehouse information (name and phone#) when we delete item #4512?  Update anomaly: Must update more than one row when one piece of data changes. Examples - What happens if the telephone number at the Sacramento North warehouse changes? What happens if the date the purchase order was placed is entered incorrectly and must be updated?

How to create a good database design?  Be an incredibly adept, intuitive database designer.  Be a fairly decent intuitive database designer and check your work with the principles of the normalization process.  Be able to follow the normalization process closely and completely. 7

8 What is normalization?  Normalization is a formal, process-oriented approach to data modeling.  Normalization is the process of: examining groups of data attributes; splitting them into appropriate entities; identifying the relationships between the entities; and identifying appropriate primary and foreign keys.

Same old/same old  Normalization should sound like what you have already done during database design.  The ultimate goals of design have not changed; we are just going to go about it in a slightly different way.  Let’s start with an application and doing it through “intuitive” database design. Student information grade screen design exercise available on the class website. 9

Database Normalization  What will you know about database normalization? Define normalization. Know the vocabulary of normalization. Understand the process of normalization. Better understand the characteristics of an effective database design.  What will you be able to do? Be able to identify the characteristics of each normal form. Be able to tell whether or not a data model is in third normal form. Potentially be able to use normalization to assist you in the design of a database. 10

11 Normalization process  Some refer to this as the “bottom-up” form of database design.  Contrast with the more intuitive “top-down” approach we have been using.  The results from the normalization process are stable, flexible entities. The results from the intuitive approach should be the same.

12 Two methods of applying normalization 1. Use it to help in designing a database. Normalization starts with a single entity. Normalization breaks that entity into a series of additional entities. More entities are discovered and named during the process. Entities are linked during the process. 2. Use it to validate the design of a database. Identify entities from the meaning of the data. Create conceptual and logical data models. Apply the rules of normalization to ensure a stable, non-redundant design.

13 Vocabulary for normalization  A “functional dependency" is a relationship between attributes in which one attribute or group of attributes determines the value of another.  A “determinant” is an attribute or group of attributes that, once known, can determine the value of another attribute.

Examples of functional dependencies and determinants  A social security number determines your name and address. SSN  name, address.  A vehicle id number determines the make and model of a car. VIN  make, model.  Name and address are “functionally dependent” on SSN.  SSN “determines” name and address.

Functional dependencies  Functional dependency diagram format: CourseID  CourseName, CourseDescription, CourseCredits ZipCode  City, State PatientID, TreatmentDateTime  TestResults  Must make decisions about the data. Are course credits dependent on the courseID? What else could course credits be dependent on? 15

16 Normalization process  Normalization is accomplished in stages. A “normal form” is a state (level of completeness) of a data model.  Unnormalized data: A data model that has not been normalized. It contains repeating groups and is not a stable model.  Unnormalized data is essentially one entity. The system under analysis is categorized as a single entity.

Steps/forms/phases in Normalization  First normal form: Remove repeating groups.  Second normal form: Remove partial functional dependencies.  Third normal form: Remove transitive dependences  There are additional normal forms. See the appendix in your DB book.  The goal for most designs is to have a database in third normal form. 17

Semester Year Student Name Student Address Student City Student State Student Zip Code Student ID Student College Student Major Student Minor Student Year Course ID Course Title Course Instructor Course Credits Grade What attributes might be needed that aren’t visible on the grade report? Group all attributes in one “big” entity. Identify a primary key for the entity. Maybe studentID for this one. Unnormalized data for grade report exercise

19 First Normal Form  First normal form: Remove repeating groups. A repeating group is an attribute or group of attributes that can have more than one value for an instance of an entity. If it is a single attribute, we have been calling it a “multi-valued” attribute.  To get a data model into first normal form: Identify repeating groups and place them as separate entities in the model. Identify a primary key for the repeating group. The key may be concatenated. Create the relationships between entities. Divide m:n relationships with appropriate intersection entities.

Second Normal Form  Second normal form: Remove partial functional dependencies.  A partial functional dependency is a situation in which one or more non-key attributes are functionally dependent on part, but not all, of the primary key. Partial functional dependencies occur only with concatenated keys.  Examples of partial functional dependencies: PatientID, TreatmentDateTime  PatName, TstResults, TrtID, LocID CourseID, StudentID  CourseTitle, Grade  Which entities developed during the transition to first normal form for the grade report have concatenated keys? 20

21 Third normal form  Third normal form: Remove transitive dependencies. A transitive dependency occurs when a non-key attribute is functionally dependent on one or more non-key attributes.  Third normal form examines entities with single primary keys and removes the “floating” or transitive dependencies.  It may be possible to have attributes that are determined by other attributes, rather than by the primary key. They must be removed into entities with appropriate primary keys.  Example of transitive dependency: courseID->course title, departmentID, department name

22 Summary of normalization process  Examine and evaluate the logical data model for effectiveness. Find the repeating groups and put the model into first normal form. Identify primary key fields for any new entities. Relate entities with foreign keys. Find the functional dependencies. Identify the partial functional dependencies and put the model into second normal form. Identify primary key fields for any new entities. Relate entities with foreign keys. Find the transitive dependencies and put the model into third normal form. Identify primary key fields for any new entities. Relate entities with foreign keys.

23 Goal of normalization A set of entities where each attribute in each entity is dependent on the primary key, the whole primary key, and nothing but the primary key.