Data Integrity: the Row Spanning Dependency antipattern.

We start out with a brief discussion of Data Integrity as applied to databases. Quick, what role does data normalization play in Data Integriry? We know what to do when an attribute has a dependency on another, non-key Attribute. But what if that dependency is to a column of a different row in the table? This happens when an unbroken sequence of values (mostly datetime) is maintained with each row representing one interval of the sequence. Look for columns with names like "FROM_xxx" and "TO_xxx". This leads to data manipulation anomolies (fancy word that means *the data integrity cannot be maintained") that cannot be fixed through normalization. Fortunately, it can be fixed and surprisingly easily. This session is encouraged for anyone signing up for "Data Modeling: Implementing Bi-Temporal Access to Data",