object relational database

An object-relational database integrates the concepts of a classical relational database with principles whose origin was taken from the object-oriented approach to software systems. The goal of this synthesis is to extend the mature technology of relational databases to the organization of arbitrary data. Object-relational databases support data objects and operations of higher complexity. Examples of complex objects are text and image data, geographical maps, data in the audio domain or databases in connection with search engines. Object-relational database models can work much faster and more efficiently in this respect because their links are in object form. Access to the object-relational database is controlled by transactions.

Object-relational database systems were realized with the background of the increased demands on the organization of large amounts of data - this is also the disadvantage of the classical relational database model. A disadvantage in particular is the limited type system of relational databases with regard to extensive texts, Excel tables, XML documents, or media data, which either cannot be stored in a relational database at all or only with increased effort. Further difficulties also arise when mapping real world objects to the relational model, this is also called" impedance mismatch".

The organization of large amounts of data

The object-relational database model integrates the model of relations, which is oriented towards simply structured data sets, with the more complex structures of an object model. Relations are effectively a collection of tables in which data can be arranged. In the object-relational model, data is stored as relations of complex objects. Again over references objects can be concatenated and also relations can stand in subset relations to each other. Thus it is expressed that a relation can be the Subrelation of another. In this respect with an object-relational data base appropriate operations are needed, which can work on also hierarchies of relations; for example around type-specifically methods of the complex objects to call. Thus it becomes clear that with the increased structural complexity of the data also the requirement for a higher functional complexity of the object-relational data base is connected.

Object-relational data model for the database structure

For the representation of complex facts, an object-relational data model must therefore provide appropriate concepts and rules for describing the structure of the database. The basis for this are the elementary data types of the relational database, on which constructive data types for complex data are defined by so-called type constructors. Type constructors are used among other things for tuples, set, lists, multi-sets. A central concept of object-relational data models are structured types. This is similar to an object type, which defines the properties (or attributes) and the methods (or behavior) of similar objects in the database. Structured types meet the requirements of object orientation, for example, in terms of structuring the data, inheritance of attributes and methods, or the principle of encapsulation. The distinction between identical and like database objects as well as the unique, unchanging object reference( OID) is guaranteed by the separation of object identity and object value. The object-oriented concept of the class is mapped in object-relational databases with so-called typed tables, which in turn are based on structured types and serve to store the data. A user-defined structured type defines the row type of this table. User-defined data types extend the data model in object-relational databases with application-specific data types.

Object-relational databases were developed with the goal of using the advantages of object-oriented databases regarding the storage of complex objects, but at the same time still being able to use the possibilities of the introduced query languageSQL. For this the existing relational model of different manufacturers was supplemented by appropriate object-oriented concepts in such a way that existing developments could be used. Thus the original data base system is made available with additional extensions, which are called for example by Informix as DataBlades or by Oracle as Cartridges. Other current systems are IBM DB2, Intersystems Cache or PostgreSQL.