## The Use of Information Capacity in Schema Integration and Translation (1993)

Venue: | In VLDB |

Citations: | 78 - 10 self |

### BibTeX

@INPROCEEDINGS{Miller93theuse,

author = {R. J. Miller and Y. E. Ioannidis and R. Ramakrishnan},

title = {The Use of Information Capacity in Schema Integration and Translation},

booktitle = {In VLDB},

year = {1993},

pages = {120--133}

}

### Years of Citing Articles

### OpenURL

### Abstract

In this paper, we carefully explore the assumptions behind using information capacity equivalence as a measure of correctness for judging transformed schemas in schema integration and translation methodologies. We present a classification of common integration and translation tasks based on their operational goals and derive from them the relative information capacity requirements of the original and transformed schemas. We show that for many tasks, information capacity equivalence of the schemas is not strictly required. Based on this, we present a new definition of correctness that reflects each undertaken task. We then examine existing methodologies and show how anomalies can arise when using those that do not meet the proposed correctness criteria. 1 Introduction Formal work on schema equivalence has largely been ignored within practical schema integration and translation tools. Practitioners have felt that theoretical work is too narrow in scope to be applicable to the problems ...

### Citations

555 | A comparative analysis of methodologies for database schema integration - Batini, Lenzerini, et al. - 1986 |

150 | View definition and generalization for database integration in a multidatabase system - Dayal, Hwang - 1984 |

144 | A theory of attribute equivalence in databases with application to schema integration - Larson, Navathe, et al. - 1989 |

117 | Superviews: Virtual integration of multiple databases - Motro - 1987 |

94 |
Relative Information Capacity of Simple Relational Database Schemata
- Hull
- 1986
(Show Context)
Citation Context ...hey can simply be an infinite list of pairs of schema instances. Clearly, such mappings are of little use in a practical context. Various restricted classes of mappings have been proposed in the past =-=[Hul86]-=-, for example, internal mappings, which only reorganize and do not invent values, and mappings that are queries in some query language. While such mappings have many desirable properties, many of them... |

76 |
A Methodology for Data Schema Integration in the Entity Relationship Model
- Batini, Lenzerini
- 1984
(Show Context)
Citation Context ... by changing one of the structures. Methodologies based on the ER model typically have transformations to change attributes to entities (and vice versa) and entities to relationships (and vice versa) =-=[BL84]-=-. Methodologies using richer models including specialization or generalization have a greater array of transformations to handle the greater number of possible mismatches [SZ91, WE79]. 5.3.1 Descripti... |

74 | Theoretical Aspects of Schema Merging
- Buneman, Davidson, et al.
- 1992
(Show Context)
Citation Context ...ns create equivalent schemas. This same example can also be modified to show that the information ordering used in other schema merging proposals does not correspond to information capacity dominance =-=[BDK92]-=-. 5.2 Merging Object Classes An alternative strategy to unioning schemas and then performing restructuring transformations is to superimpose structures that have been identified as referring to the sa... |

66 | Schema Equivalence in Heterogeneous Systems: Bridging Theory and Practice
- Miller, Ioannidis, et al.
- 1994
(Show Context)
Citation Context ...yzing and presenting the transformations, we use a single data model for describing the relevant schema families. Specifically, we employ schema intension graphs (SIGs), a formalism defined elsewhere =-=[MIR]-=-. Within SIGs, constraints that impact the information capacity of a schema are explicitly represented in a simple graph notation. These graphs are therefore a convenient tool for proving or disprovin... |

57 |
The category concept: An extension to the entity-relationship model
- Elmasri, Weeldreyer, et al.
- 1985
(Show Context)
Citation Context ..., the transformed schema must have equivalent information capacity to the original schemas (Section 3). The methodology uses the entity-category-relationship (ECR) model, an extension to the ER model =-=[EHW85]-=-. 5.2.2 Description of Transformation We describe a merge transformation for entity classes with a single attribute as key. We present only one alternative of a family of transformations from this met... |

56 | Integrating user views in database design - Navathe, Elmasri, et al. - 1986 |

53 | The format model: a theory of database organization - Hull, Yap - 1984 |

46 |
Antonelles. Relational database theory
- Atzeni, D
- 1993
(Show Context)
Citation Context ...cond constraint corresponds to the relationship uniqueness assumption (RUA), often asserted in relational theory, that a schema represents at most one relationship between any given set of attributes =-=[AP82]. Un-=-der this assumption, and the assumption that the keys of the object classes are identical, the map f is an information preserving mapping and S1 �� S2. Now consider the inverse mapping function g ... |

44 |
A Formal View Integration Method
- Biskup, Convent
- 1986
(Show Context)
Citation Context ...the RUA then many of the transformations also guarantee dominance in the opposite direction. Similar transformations are included in methodologies based on the relational model to merge two relations =-=[BC86]-=-. In that work, the precondition for applying the transformation is the existence of an integration constraint asserting that the projection of the two relations on the common attributes (including th... |

39 |
A.: On the correctness of representing extended entityrelationship structures in the relational model
- Markowitz, Shoshani
- 1989
(Show Context)
Citation Context ...traint is sufficient to ensure that the transformed schema has equivalent information capacity to the original. Other relational work has used less restrictive preconditions for merging two relations =-=[MS92]-=-. Null dependencies and general inclusion dependencies are used to ensure information capacity is preserved. 5.3 Structural Mismatch The merging of schemas (whether via unioning or superimposition) is... |

29 | Constructing Superviews - Motro, Buneman - 1981 |

24 |
Binary-relationship integration methodology. Data and Knowledge Engineering
- Shoval, Zohn
- 1991
(Show Context)
Citation Context ...ray of transformations to handle the greater number of possible mismatches [SZ91, WE79]. 5.3.1 Description of Methodology We focus on a view integration methodology included in a database design tool =-=[SZ91]-=-. The tool recognizes structural differences and either applies a single resolution transformation or provides the designer with a choice of resolution transformations. We examine two situations and s... |

23 |
Methods and Tools for Equivalent Data Model Mapping Construction
- Kalinichenko
- 1990
(Show Context)
Citation Context ...stance mappings that these induce. In the next section, we show that by ensuring that such mappings 2 Note that this condition implies the data model commutative mapping principle described elsewhere =-=[Kal90]-=- and corresponds to the definition of lossless schema transformations [Tro93]. are in fact information capacity preserving some very important operational goals can be achieved. 3 Taxonomy of Schema I... |

22 |
On Data Schema Transformations
- Troyer
- 1993
(Show Context)
Citation Context ...ing that such mappings 2 Note that this condition implies the data model commutative mapping principle described elsewhere [Kal90] and corresponds to the definition of lossless schema transformations =-=[Tro93]-=-. are in fact information capacity preserving some very important operational goals can be achieved. 3 Taxonomy of Schema Integration and Translation Tasks Schema translation and integration are neces... |

18 | Theoretically sound transformations for practical database design - Rosenthal, Reiner - 1988 |

16 | Inclusion and Equivalence between Relational Database Schemata - Atzeni, Ausiello, et al. - 1982 |

16 | elsberg. Matching techniques in Global Schema Design - Mannino, E - 1984 |

13 | The structural model for database design - Wiederhold, El-Masri - 1979 |

12 | A methodology for the design and transformation of conceptual schemes - Eick - 1991 |

11 | On Equivalences of Database Schemes - Rissanen - 1982 |

10 |
Formal properties of net-based knowledge representation schemes
- Atzeni, Jr, et al.
- 1986
(Show Context)
Citation Context ...hema rather than to an instance of a schema that may hold contradictory information. Instance level conflicts are beyond the scope of this paper. 9 Other authors refer to such schemas as inconsistent =-=[AP86]-=-. It is critical that a schema transformation tool recognize potential conflicts and use this information to guide the designer in correctly specifying constraints and choosing transformations. A tool... |

5 | Data model equivalence - Borkin - 1978 |

5 |
Database design tools: Combining theory, guesswork, and user interaction
- Rosenthal, Reiner
- 1989
(Show Context)
Citation Context ...nce has largely been ignored within practical schema integration and translation tools. Practitioners have felt that theoretical work is too narrow in scope to be applicable to the problems they face =-=[RR89]-=-. As a result their work is driven by an intuitive, rather than formal, notion of correctness. Some recent work on translation and integration has successfully used information capacity equivalence as... |