Ten Things to Avoid in a Data Model
Dr. Michael Blaha Modelsoft Consulting Corp www.modelsoftcorp.com E-mail: [email protected]
Introduction
A model is an abstraction of some aspect of a problem. A data model is a model that describes how data is represented and accessed, usually for a database.
Data modeling can be a difficult task and is often pivotal to the success or failure of a project.
There are many pitfalls to data modeling as we will explain...
Strategic pitfalls. Detailed pitfalls.
We do not discuss detailed modeling constructs such as keys, data types, nullability, and referential integrity.
PAGE 2
Strategic Pitfalls
PAGE 3
Strategic Pitfall: Vague Purpose
Dont build a model without understanding the business rationale. The purpose for a model dictates the level of detail.
Just entities and relationships. Fully attributed. With data types and constraints.
The purpose also dictates the level of polish, the degree of completeness, and the amount of time justified. Different kinds of data models.
Detailed application model for development. Rough application for a purchase spec. Enterprise model for integration.
This pitfall might seem obvious, but Ive seen modeling efforts with little business purpose and no clear definition of deliverables.
PAGE 4
Strategic Pitfall: Literal Modeling
Your job is not to do what the customer says. Your job is to solve the problem that the customer is imperfectly describing. You must pay attention to the hidden true requirements. You must interpret and abstract what the customer tells you.
You must recognize arbitrary business decisions that could easily change.
You can raise abstraction by thinking in terms of patterns. The use case mentality really misses this point.
PAGE 5
Strategic Pitfall: Literal Modeling Example
Original literal model
Improved abstract model
The original model is correct, but has problems. What happens if a person gets promoted to a supervisor and then to a manager? Are there multiple records? Movement of a record? Or??? The improved model is more abstract and softcodes the management hierarchy.
PAGE 6
Strategic Pitfall: Large Size
Avoid large models. Limit a model to no more than 200 tables. Large models involve more work. Is the large size really justified or can you simplify the model with abstraction? I rarely encounter a large model with a compelling justification. I dont see this step in software development methodologies, but it is certainly needed.
PAGE 7
Strategic Pitfall: Speculative Content
Do not include content that is not needed now and might be helpful in the future.. All this does is to make a model larger, increase development time, and raise cost. A model must fully address the requirements, but not greatly exceed them. A quality model should be readily extended, so there is no need to add content in advance of need. Speculative content runs counter to the philosophy of agile development.
PAGE 8
Strategic Pitfall: Lack of Clarity
A relational database is declarative. Declare data in your models. A domain is the set of possible values for an attribute.
ERwin lets you define domains and then assign them to the pertinent attributes.
An enumeration is a domain that has a finite set of values.
Declare enumerations in your databases.
Dont store data structures with a binary encoding. Dont use cryptic names. Dont use anonymous fields that application code must interpret. Obfuscation can happen through sloppy development practices.
PAGE 9
Strategic Pitfall: Lack of Clarity Example
Car table
Enumeration stored in place
carID year 1 2 3
color weight 2000 1500 2500
2001 red 1989 red 2000 blue
Car table
Color table
colorID 1 2 3 color red green blue 2000 1500 2500
Enumeration stored separately
carID year 1 2 3
colorID weight
2001 1 1989 1 2000 3
Car table
Enumeration encoded
carID year 1 2 3
color weight 2000 1500 2500
2001 1 1989 1 2000 3
PAGE 10
Detailed Pitfalls
PAGE 11
Detailed Pitfall: Reckless Violation of Normal Forms
Do not accidentally violate normal forms. A normal form is a guideline that increases data consistency. As tables satisfy higher levels of normal forms, they are less likely to store redundant or contradictory data. Denormalization is only justified when there is a major performance bottleneck, such as for data warehouses. Be suspicious of large tables (30 attributes or more). Be suspicious of any entity type that is difficult to define. It is acceptable to violate normal forms deliberately, when there is a good reason to do so.
PAGE 12
Detailed Pitfall: Normal Forms Example
Violates normal form
Satisfies normal form
The contact position and contact phone depend on the contact name. The contact name depends on customerPK.
PAGE 13
Detailed Pitfall: Needless Redundancy
Be careful with redundancy.
Redundancy across applications. Redundancy within an application.
Normal forms are one aspect of redundancy. Ideally there should be a single recording of each data item. (Rarely is this completely feasible.) Organizations are rife with applications that overlap in awkward and loosely controlled ways.
This is a major justification for data warehouses.
Dont include redundant data to compensate for a poorly conceived application. Redundant data is acceptable if you use built-in database features to keep redundant data consistent (such as materialized views).
PAGE 14
Detailed Pitfall: Parallel Attributes
Avoid parallel attributes for non-data-warehouse applications. Parallel attributes often codify arbitrary business decisions, reducing information system flexibility.
Parallel attributes
Parameterized model
Widespread use of parallel attributes often indicates a poor model.
PAGE 15
Detailed Pitfall: Symmetric Relationships
Avoid symmetric relationships for relational databases. Promote a symmetric relationship to an entity type.
Symmetric relationship
Promotion to an entity type
Otherwise double entry or double search. Symmetric relationships can be acceptable for programming.
PAGE 16
Detailed Pitfall: Anonymous Fields
As much as possible, clearly describe the data being stored and avoid anonymous fields.
fragment of Location table
locationAddress1 456 Chicago Street 198 Broadway Dr. 123 Main Street Chicago, IL xxxxx locationAddress2 Decatur, IL xxxxx Suite 201 Cairo, IL xxxxx locationAddress3 Chicago, IL xxxxx
How to distinguish the city of Chicago from Chicago street? May need to parse a field to separate city, state, and postal code. A few incidental user-defined fields are OK.
PAGE 17
Summary
Data modeling is often a pivotal task in building a database application. A data model determines an applications data quality, extensibility, and performance and influences whether the application has a chance at business success. You can improve your data models if you pay attention to the pitfalls we have covered.
PAGE 18
Speaker Bio
Since 1994 Dr. Michael Blaha has been a consultant and trainer in conceiving, architecting, modeling, designing, and tuning databases for dozens of organizations throughout the world. He has authored six U.S. patents, five widely used books, and many papers. His most recent book, Patterns of Data Modeling, was published in June 2010. Blaha received his doctorate from Washington University in St. Louis and is an alumnus of GE Global Research in Schenectady, NY. You can contact him at [email protected] and www.modelsoftcorp.com.
PAGE 19
Questions?
PAGE 20
