Dimensions consist of levels, hierarchies, and level attributes. The logical model provided by the OLAP Catalog consists of the following entities: ▪ĭimensions: Dimensions are used to express relationships, such as hierarchies, in your data. We use these notations throughout the book. Commercial relational databases support both semantics. Unless we state otherwise, our discussion will assume set semantics. In multi-set semantics, a tuple can appear any number of times in each relation, and hence a database instance is an assignment of a multi-set of tuples to each relation. That is, a tuple can only appear once in a relation. In set semantics, a database state assigns a set of tuples to each relation in the database. We distinguish between set semantics of databases and multi-set semantics. We can test explicitly for NULL with the predicate is NULL.Ī state of the database, or database instance, is a particular snapshot of the contents of the database. This makes intuitive sense because if two values are not known, we certainly do not know that they are equal to each other. The important property to keep in mind about the NULL value is that an equality test involving a NULL returns NULL.
For example, the value of an age attribute may be NULL if it's not known, whereas the value of spouse could be either unknown or may not exist. Intuitively, NULL means that the value is not known or may not exist. We pay special attention to the NULL value in a database. In some cases, we describe a tuple with a mapping from attribute names to values, such as: Interview ( Alan Jones,, Annette Young, No, 2.8) If the relation being discussed is clear from the context, we denote its tuples with its values in parentheses, e.g., A tuple assigns a value to each attribute of the table. One of the challenges we face in data integration is that different sources organize their data differently, and these differences need to be reconciled.Ī relational table includes a finite number of rows, called tuples (or records). The EmployeePerformance table only provides a single grade, and does not model the fact that this grade may be the composition of several more specific performance measures.
For example, the Interview table does not include an attribute for the location of the interview or the position for which the candidate was interviewing. In a sense, the schema describes how the database author decided to organize the data, what aspects of the data she chose to model, and the distinctions she wished to make.