Browse Prior Art Database

Simplification of Queries in Relational Data Bases

IP.com Disclosure Number: IPCOM000085383D
Original Publication Date: 1976-Mar-01
Included in the Prior Art Database: 2005-Mar-02
Document File: 3 page(s) / 31K

Publishing Venue

IBM

Related People

Chandra, AK: AUTHOR

Abstract

Given a relational model of a data base [1] with which a user would interact via some user interface (user's model) it is suggested (Fig. 1) that queries be translated into an intermediate model in which there are only binary relations. These intermediate relations may well be different from the relations in the users model.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 3

Simplification of Queries in Relational Data Bases

Given a relational model of a data base [1] with which a user would interact via some user interface (user's model) it is suggested (Fig. 1) that queries be translated into an intermediate model in which there are only binary relations. These intermediate relations may well be different from the relations in the users model.

The implementation would then have to handle only binary relations, which seems easier than handling relations of arbitrary rank - e.g., [2] suggests how the implementation of queries on binary relations can be speeded up. Also the intermediate model is itself a relational model, so that queries in it are amenable to very high-level optimization as well as lower level optimization.

It is not necessary that the data structure used by the implementation explicitly have each binary relation of the intermediate model in an array - any data structure in which the relations of the intermediate model can be computed, is adequate, for example, an array for each relation in the user's model is adequate.

There is described herein the intermediate model, and how queries in the user's model can be translated into it. (See Fig. 2).

Relations of ranks 1, 2 remain unchanged in the intermediate model (strictly, unary relations can be thought of as binary relations where the element is repeated). Suppose the remaining relations are called, R(1), ..., R(n), and have ranks k(1), ..., k(n), respectively. Corresponding to each R(i) are relations R(i,1), R(i,2), ..., R(i,k(i)) in the intermediate model.

The first coordinate of each of these relations corresponding to R(i) is from any domain that indexes over the vectors in R(i). This could be the integers as in the example or the set of vectors R(i) itself. The latter is used in the description below.

Intermediate relation R(i) is defined as follows: vector (a(1)..., a(j),...,a(k(i)) is in relation R(i) if and only if ((a(1),...,a(k(i)), a(j)) is in R(i,j).

Given any query Q...