INF2603 Summary.

Data Models: - designer + programmer + end user - representation of data structure Properties of a relation: - table is 2D w/ rows and columns - row = tuple - row represents single entity occurrence w/in entity - column represents attribute w/ unique name - intersection of row and column represents single data value - values in column must have same format - column's specific range of values = attribute domain - order is immaterial - each table must have attribute or collection of attributes that uniquely identifies each row Business Rules: - unary and recursive is when entity has relationship with self (manager is employee who manages other employees) - binary is when two entities are associated - ternary and higher order relationships have associations w/ 3+ entities (conceptual is when one entity is not attached and logical is when all other entities stem from one) Relational Model: - relational is logical not physical + may promote islands of info - hierarchy is no standards, changes require all - Represents data as independent relations Object-oriented: - depicted with UML - abstraction of real world entity - attributes describe properties of objects - class resembles ER model entity set - method represents real world action Joins: - natural links by selecting only the rows with common values in their common attributes - product of tables - select performed on output (join columns) - project performed to remove duplicates - line infinity - equijoin links based on equality condition - theta join for other conditions - inner join returns matched records for tables being joined - outer joins matched records retained and unmatched null Union Compatibility: - when two+ tables share the same o. columns and columns share same/compatible domains Existence Dependent: - can exist in DB only when associated with another entity related occurrence independent = strong/regular entity