Attributes, called data type attributes (hereafter DTA), define fields with self-defined data values.In the TNDB, all attributes are divided into two disjoint classes that have fundamentally different purposes: Attributes are assigned symbolic names that can be changed.Įach row of a table in the database is uniquely identified by a pair “table number, row number”, and each field in the table is uniquely identified by a tuple “table number, row number, attribute number”. The numbers of the deleted attributes are not assigned to new ones. For ease of programming and maintenance, tables are assigned symbolic names that can be changed.Įach attribute has a unique number that never changes after it has been created. The numbers of deleted tables are not assigned to new ones. Tables have unique numbers that never change.
All rows in the same table have the same number of fields. Two concepts of relational databases - the “domain” of the database and the “attribute” of an individual relation - in essence merge into one.Įach column in each table corresponds to one of the database attributes.
Tables in the TNDB have no headers, that is, the columns do not have a special name (“attribute” in the terminology of relational model).Ī common set of attributes is defined for the entire database as characteristics that are necessary and sufficient to formalize the structure and parameter description of the simulated system.
The selection of fields with the same serial number in all rows constitutes the columns of the table. Identical rows in tables are not allowed. The TNDB consists of a set of common attributes, a set of tables, and relations between the attributes and columns of the table.Ī table in the TNDB is defined as a structure in the form of an ordered set of rows (records, tuples) that have an ordered set of fields (cells). The tabular network model itself and the technology used for its implementation are called the tabular network database (hereafter TNDB).įigure 1: Abstract structure of a tabular network database. The network “component” of the data model allows in many cases to eliminate the costs of clustering, indexing and data search. It simplifies the adequate modelling of the real world, allows reducing the number of errors and alterations in the design and maintenance of the database. The preservation of knowledge about the relationships of entities in an explicit form and their effective use is an important advantage of the tabular network model according to the criterion of simplicity and clarity. This is typical for “one-to-many” and “many-to-many” relationships, when additional, so-called “connecting”, tables are used for their formalization. In the relational model, the objective knowledge about the relationships of real entities that are relevant during creation and replenishment of a database is lost as evident and is transformed by developers into indirect one.
This model is very simple at the conceptual level - its essence is that not only the actual data can be stored in the fields of tables, but also direct address links to rows of any tables. The proposed data model is called “tabular network” because tables form the basis for structuring information and relationships are formalized in an explicit form, as in earlier network databases.
Tabular network database - new approach to building data models and related software at the intersection of the concepts of relational and network databases and object-oriented programming