Lecture
As for the general terminology of the relational approach, we will actively use the relevant terms. These terms include the names of relational operations - selection, projection, connection; names of set-theoretic operations - union, intersection, difference, etc.
In cases where the traditional terminology of System R is at odds with the generally accepted one, we will prefer the terminology of System R. In particular, this concerns the use of the term “relationship field” instead of “relationship attribute”.
In the System R itself, when switching to commercial systems, there was also a certain change in terminology. In particular, in some recent publications there has been a tendency to use more familiar terms among IBM users: file, record, etc. We will use the terms System R, more closely related to relational systems. Next, we will describe some of the basic terms of System R, based primarily on theoretical considerations, and seeking to reflect the practical aspects of the relevant concepts.
The basic concept of System R is the concept of a table (an attitude close to realization of the basic concept of a relational approach; sometimes, depending on the context, we will use this term). A table is a regular structure consisting of a finite set of similar entries — tuples. Each tuple of one relation consists of a finite (and the same) number of fields of a tuple, with the i-th field of each tuple of one relation containing only one type of data, and the set of valid data types in System R is predefined and fixed. By virtue of the regularity of the relationship structure, the concept of a tuple field is expanded to the concept of a table field. The i-th table field can be interpreted as a set of single tuples obtained by sampling the i-th fields from each tuple of this table, i.e. in the standard terminology as a projection of the relationship on the i-th attribute. System R terminology does not include the concept of a domain, it is replaced here by the concept of a field type, i.e. the type of data that is allowed in this field (this is not a completely equivalent replacement, but that is the reality of System R).
The tables that make up the System R database can be physically stored in one or more segments, which are easiest to understand as external memory files (and this is quite true). Segments are divided into pages, in which relations tuples and auxiliary service structures of data indices are located. Accordingly, each segment contains two groups of pages - data pages and pages of index information. Pages of each group have a fixed size, but pages with index information are smaller in size than data pages. Data pages can contain tuples of more than one relation (this is a very important property of the physical organization of the System R databases; the advantages resulting from this organization will be explained later).
This, of course, does not exhaust the set of System R concepts, but we will explain the rest of the terms in the course of the exposition, since this requires an appropriate conceptual context.
Comments
To leave a comment
IBM System R — реляционная СУБД
Terms: IBM System R — реляционная СУБД