Showing papers by "Katsumi Tanaka published in 1979"
06 Nov 1979
TL;DR: The semantic aspects of functional and multivalued dependencies are discussed in order to choose a better logical design of a relational database and a mixed design approach of a conceptual and a logical designs is introduced.
Abstract: In this paper, we mainly discuss the semantic aspects of functional and multivalued dependencies in order to choose a better logical design of a relational database. We clarify the differences between a conceptual design and a logical design of a relational database. For example, a multivalued dependency does not always capture a conceptual dependency such that a data element semantically determines a set of other data elements. Moreover, some transitively specified multivalued dependencies are shown to often impose a semantically unnatural constraint. A sufficient condition for these multivalued dependencies to hold in a natural sense is provided. A mixed design approach of a conceptual and a logical designs is introduced, in which each design approach is compensated for its deficiency by the other one. Some generalization of a relational model is also suggested to handle some interface problems between a logical design and a conceptual design, such as a null value problem and an entity identification problem.
10 citations
TL;DR: Graph theoretic properties of CR files and QCR files are discussed by using the properties of interval graphs well known in graph theory by providing a basic condition for the existence of a QCR file without redundancy for a given buffer size.
9 citations
Journal Article•
7 citations
06 Nov 1979
TL;DR: Functional dependencies, which play an important role in relational database design, and time-dependent functional dependencies are pointed out to be useful in processing retrieval/update operations.
Abstract: In the database field, the requirement of high level facilities for retrieval/update operations is now increasing rapidly. Our approach from the relational database point of view for contribution to this problem is to provide 1) efficient processing of relational retrieval/ update operations, and 2) a high level user interface. In order to achieve this goal, a new concept concerning "abstracted characteristics" is presented. Abstracted characteristics are defined to be characteristics abstracted from sets of tuples in relations stored in the database. A classification of abstracted characteristics is presented. Functional dependencies, which play an important role in relational database design, and time-dependent functional dependencies are pointed out to be useful in processing retrieval/update operations. Some important applications of abstracted characteristics are discussed. Among them: 1) efficient processing of retrieval/update operations, 2) powerful view update checking facilities, 3) providing some rough meanings of null responses and 4) a high level user interface.
3 citations