90 likes | 218 Vues
In this class agenda from September 18, 2012, we discuss various aspects of database design and evaluation techniques focused on preparing for the exam. Key topics include defining rules of thumb for effective database design, evaluating each other’s homework (HW#4), and understanding design heuristics. We emphasize examining 1:1 relationships, logical placement of foreign keys, and utilizing design evaluation criteria, such as entity attributes and primary keys, to identify potential issues. Collaborative group work is encouraged for peer evaluation of database designs.
E N D
Class Agenda – 09/18/2012 Answer questions about the exam. Define rules of thumb for database design. Evaluate each other’s database design for HW#4.
Examine all 1:1 relationships. Determine whether the attributes could all be placed in just one of the two entities eliminating the need for one of the two entities. • In a 1:1 relationship, the foreign key can be placed in either entity. Put the foreign key in the entity that seems most logical, has the fewest number of rows and/or also will result in the fewest number of null values.
Evaluate design homework • Break into groups • Exchange assignments • Put your name (as the “evaluator”) on the back of the assignment • Evaluate the syntax and logic of the design
“Syntax” Database Modeling Issues Does the diagram contain entities, attributes, primary keys and relationships? Is it structured using the modeling format discussed in this class? Does each entity have a primary key? Is the primary key labeled as a primary key? Is each relationship labeled with a verb or verb phrase? Does each relationship have a foreign key? Is the foreign key labeled as a foreign key?
“Logic” Database Modeling Issues • Is all necessary data represented on the model? What’s missing? • Is the data grouped correctly with the appropriate entity? • Does each entity have a primary key that will yield a unique value for each row? • Are the entities related accurately? • Are the cardinalities of the relationships accurate? • Are there any M:N relationships? • Will the design yield redundant data?