Database Repairing and Consistent Query Answering

By L. Bertossi. 2011

341 citations recorded by the source corpus.

Integrity constraints are semantic conditions that a database should satisfy in order to be an appropriate model of external reality. In practice, and for many reasons, a database may not satisfy those integrity constraints, and for that reason it is said to be inconsistent. However, and most likely, a large portion of the database is still semantically correct, in a sense that has to be made precise. After having provided a formal characterization of consistent data in an inconsistent database, the natural problem emerges of extracting that semantically correct data, as query answers. The consistent data in an inconsistent database is usually characterized as the data that persists across all the database instances that are consistent and minimally differ from the inconsistent instance. Those are the so-called repairs of the database. In particular, the consistent answers to a query posed to the inconsistent database are those answers that can be simultaneously obtained from all the database repairs. As expected, the notion of repair requires an adequate notion of distance that allows for the comparison of databases with respect to how much they differ from the inconsistent instance. On this basis, the minimality condition on repairs can be properly formulated. In this monograph we present and discuss these fundamental concepts, different repair semantics, algorithms for computing consistent answers to queries, and also complexity-theoretic results related to the computation of repairs and doing consistent query answering. Table of Contents: Introduction / The Notions of Repair and Consistent Answer / Tractable CQA and Query Rewriting / Logically Specifying Repairs / Decision Problems in CQA: Complexity and Algorithms / Repairs and Data Cleaning

Subfields: Artificial Intelligence, Computer Networks and Communications

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Sanjay Krishnan
2. Ihab F. Ilyas
3. Eugene Wu
4. Xu Chu
5. Ken Goldberg
6. M. Franklin
7. Paolo Papotti
8. Jiannan Wang

In Computer Networks and Communications

1. Xu Chu
2. Ihab F. Ilyas
3. Paolo Papotti
4. Sanjay Krishnan
5. Eugene Wu
6. Ken Goldberg
7. M. Franklin
8. M. Tamer Özsu

Citations in the corpus

• AlphaClean: Automatic Generation of Data Cleaning Pipelines (2019, background): “A large body of literature addresses relational queries over databases with errors by focusing on specific classes of queries [5], leveraging constraints over the input relation [10], integration with crowdsourcing [8].”
• BoostClean: Automated Error Detection and Repair for Machine Learning (2017, background): “Similarly, work on the consistent query answering problem explored the minimal effort needed to answer a query given a set of integrity constraints over a dirty relation [11].”
• Trends in Cleaning Relational Data: Consistency and Deduplication (2015, methodology): “We refer readers to a survey [8] for a comprehensive treatment of consistent query answering.”
• Discovering Denial Constraints (2013, methodology): “DCs serve as a great compromise between expressiveness and complexity for the following reasons: (1) they are defined on predicates that can be easily expressed in SQL queries for consistency checking; (2) they have been proven to be a useful language for data cleaning in many aspects, such as data ”