Showing papers on "Tuple published in 1976"

Decomposition—a strategy for query processing

Abstract: Strategy for processing multivariable queries in the database management system INGRES is considered. The general procedure is to decompose the query into a sequence of one-variable queries by alternating between (a) reduction: breaking off components of the query which are joined to it by a single variable, and (b) tuple substitution: substituting for one of the variables a tuple at a time. Algorithms for reduction and for choosing the variable to be substituted are given. In most cases the latter decision depends on estimation of costs; heuristic procedures for making such estimates are outlined.

The design of a rotating associative memory for relational database applications

Abstract: The design and motivation for a rotating associative relational store (RARES) is described. RARES is designed to enhance the performance of an optimizing relational query interface by supporting important high level optimization techniques. In particular, it can perform tuple selection operations at the storage device and also can provide a mechanism for efficient sorting. Like other designs for rotating associative stores, RARES contains search logic which is attached to the heads of a rotating head-per-track storage device. RARES is distinct from other designs in that it utilizes a novel “orthogonal” storage layout. This layout allows a high output rate of selected tuples even when a sort order in the stored relation must be preserved. As in certain other designs, RARES can usually output a tuple as soon as it is found to satisfy the selection criteria. However, relative to these designs, the orthogonal layout allows an order of magnitude reduction in the capacity of storage local to the search logic.

High level domain definition in a relational date base system

Abstract: A relational data base is herein defined as a collection of normalized relations (relations in first normal form) and a collection of domains. A normalized relation may be viewed as a table, wherein each row of the table corresponds to a tuple of the relation, and the entries in a given column belong to the set of values constituting the domain underlying that column. The domains of a data base have an abstract existence apart from the data base relations.The data base also includes various types of semantic integrity rules, which specify additional properties of the data in the data base. One such type of semantic integrity rule is the domain definition. A domain definition includes the precise description of the set of values (objects) constituting the domain. In a normalized data base, all domains are sets of atomic data values. A domain definition also includes a specification of the ordering on the values in a domain, for comparability purposes. In addition, a domain definition contains a specification of the action that is to occur if an attempt is made to violate the restriction that every entry in each column of a relation must be from the underlying domain of that column.A nonprocedural language permitting the high level expression of domain definitions is defined. Language details and examples are presented, and the syntax and informal semantics of the domain definition language are given. This approach to domain definition is analyzed in terms of its impact on other aspects of data base semantic integrity. The relationship with the data base system in general is outlined. An analysis of intradomain and interdomain comparability is included. An introduction to relevant implementation issues is also presented. Emphasis is placed on a general approach to implementation and implementation techniques, rather than on a specific system.

User Extensions to the Peterlee Relational Test Vehicle

Abstract: In the design of the PRTV system we stressed the importance of the easy writing of extensions to the language, ISBL. We give sufficient details of the ISBL language to provide a framework in which to discuss these extensions. They fall into two classes, those which can be thought of as relations, computed when required, rather than being explicitly stored, and those where a tuple in the result relation is, in general, formed from values taken from more than one tuple in the input relation. We discuss these extensions at a theoretical level, how they are defined and used, and how they are implemented. Finally we discuss the operation of 'glumping' or partitioning, in terms of relations not in first normal form. This provides a general way of lookinq at such operations as subtotalling.

An attribute-based file organization for a relational database

Abstract: A file organization based on the work of Wong and Chiang (12) may be used to store a relational database which comprises several relations. In the proposed storage structure, keywords are defined on domains (other than primary key domains) and used to divide the file into disjoint subsets called Boolean atoms. Tuples in each Boolean atom are stored contiguously and identifiers for each atom are stored in a directory. The actual construction of the atom directory and the storage of the tuples are presented with an example. Procedures for processing queries against the proposed structure are illustrated using sample queries from SQUARE (1). A modification of the structure to accommodate ranges of values as keywords is also indicated.

Improved updating in relational data base systems by deuter-sphere algorithms

Abstract: The need to ease the user's handling of data banks concerning natural insertions and deletions, requires more data independencies. A fundamental trend to 'normalize' relations in first and second normal forms has been indicated by Codd and Kent. When it comes to the third normal form, obscurities are spreading. This touches the very moment, when adequate content relations of programming morphologies to content realities are missing. As Chamberlin, Gray and Traiger pointed out, a so called 'rectangle of a base relation' aids to clarify natural function dependencies. These indications have been further developed and advanced to a level, where each tuple and entity is subdivided into six deuter-criteria: 1. identity, 2. age, 3. association, 4. frequency, 5. significance and 6. truth. The validity of each criterion is related to all potential components within a relational data base. The metric data of each deuter criterion can be derived by analytical methods within three- and four dimensional realities in the model of a deuter-sphere. This leads to fundamental insertion- and deletion rules and an improved updating.