Computer World. Матросова Т.А. - 59 стр.

UptoLike

ВУЗ: 

УлГТУ | Ульяновск

Составители: 

Рубрика: 

58
database theory and database technology and increasingly sophisticated requirements
of database users. The quantitative evolution of DBMSs stems from the increasing
number and variety of database applications and the increasing diversity of
computing resources.
A DBMS is characterized mainly by the data model it supports. The first
DBMSs, based on the hierarchical or network model, remain the most used today.
They can be viewed as an extension of file systems in which interfile links are
provided through pointers. The data manipulation languages of those systems are
navigational, that is, the programmer must specify the access paths to the data by
navigating in hierarchies or networks.
In the early 1980s, the first systems based on the relational model appeared on
the market, bringing definite advantages over their predecessors. Today a large
number of relational products is available on mainframe computers, minicomputers,
microcomputers, and dedicated computers (database machines), and their market is
rapidly expanding. The success of the relational model among researchers, designers,
and users is due primarily to the simplicity and the power of its concepts.
What Is a Relational Database Management System?
The advantages of the relational model, invented by E. F. Codd, have been
thoroughly demonstrated by database researchers. One main advantage is the ability
to provide full independence between logical data descriptions, in conceptual terms,
and physical data descriptions, in terms of files.
As a consequence of this physical independency, high-level data manipulation
languages may be supported. Such languages free the programmer from physical
details, thereby allowing query optimization to be done by the system rather than by
the user. The promotion of the relational model has also been helped by database
language standardization, which yields the standard Structured Query Language
(SQL). SQL provides a uniform interface to all types of users (database
administrators, programmers, end users) for data definition, control, and
manipulation.
The relational data model can be characterized by three features:
1. The data structures are simple. These are two-dimensional tables, called
relations (or tables), whose elements are data items. A relation can be viewed as a
file; a row of a relation, called tuple (or row), can be viewed as a record; and a
column of a relation called attribute (or column), can be viewed as a data item. The
relationship linking two relations is specified by a common attribute in both relations.
For example, the relationship between an EMPLOYEE relation and a
DEPARTMENT relation can be specified by an attribute dept_name stored in both
relations.
2. A set of eight operators (union, intersection, difference, Cartesian product,
select, project, join, and difference), called relational algebra, facilitates data
definition, data retrieval, and data update. Each relational operator takes one or two
relations as input and produces one relation.
3. A set of integrity constraints defines consistent states of the database.

Страницы