Relational database

The relational model organizes information through tuples and attributes, commonly presented as rows and columns. Separate tables can describe people and events and relate them through explicit keys. Queries primarily express which information to select or combine rather than requiring each application to specify its physical location on disk.

References: [1]

Selection keeps rows meeting conditions, projection chooses columns, and joins combine tables using stated relationships. An overly broad join can generate many unintended combinations. Constraints such as foreign keys help maintain valid references, making the definition of relationships central to the meaning and correctness of the resulting information.

References: [2]

An optimizer may satisfy one query using scans, indexes, or different join orders. Table rows have no guaranteed display order unless a query requests one. Practical SQL systems also introduce duplicate rows, null values, and extensions, whose behavior must be understood alongside the underlying relational organization rather than assumed away.

References: [1]

Updates must preserve constraints as rows are inserted, changed, or removed. Transactions organize related operations into units of completion or rollback, working with concurrency control and recovery. A tabular interface alone does not establish transactional guarantees; those depend on the rules and mechanisms implemented by the database system.

References: [2]