Whereas physical database tuning has received a lot of attention over the last decade, logi- cal database tuning seems to be under-studied. We have developed a project called DBA Companion devoted to the understanding of logical database constraints from which logical database tuning can be achieved.

In this setting, two main data mining issues need to be addressed: the first one is the design of efficient algorithms for functional dependencies and inclusion dependencies inference and the second one is about the interestingness of the discovered knowledge. In this paper, we point out some relationships between database analysis and data mining. In this setting, we sketch the underlying themes of our approach. Some database applications that could benefit from our project are also described, including logical database tuning.

Foreign keys form one of the most fundamental constraints for relational databases. Since they are not always defined in existing databases, algorithms need to be devised to discover foreign keys. One of the underlying problems is known to be the inclusion dependency (IND) inference problem. In this paper a new data mining algorithm for computing unary INDs is given. From unary INDs, we also propose a levelwise algorithm to discover all remaining INDs, where candidate INDs of size i + 1 are generated from satisfied INDs of size i; (i > 0). An implementation of these algorithms has been achieved and tested against synthetic databases. Up to our knowledge, this paper is the first one to address in a comprehensive manner this data mining problem, from algorithms to experimental results.

From statistics, sampling technics were proposed and some of them were proved to be very useful in many database applications. Rather surprisingly, it seems these works never consider the preservation of data semantics. Since functional dependencies (FDs) are known to convey most of data semantics, an interesting issue would be to construct samples preserving FDs satisfied in existing relations. To cope with this issue, we propose in this paper to define Informative Armstrong Relations (IARs); a relation s is an IAR for a relation r if s is a subset of r and if FDs satisfied in s are exactly the same as FDs satisfied in r. Such a relation always exists since r is obviously an IAR for itself; moreover we shall point out that small IARs with interesting bounded sizes exist. Experiments on relations available in the KDD archive were conducted and highlight the interest of IARs to sample existing relations.

Understanding data semantics from existing relational databases is important for several applications such as database maintenance and analysis, database re-engeenering or query optimization. Data semantics is carried out in particular by integrity constraints. For most operational databases, particularly for the oldest ones, we cannot assume to dispose of this knowledge. In this paper, we present a tool called DBA Companion which can be an help to deal with the understanding of existing relational databases. This tool integrates algorithms to deal with the extraction of integrity constraints and several related problems. From this mined knowledge, the logical tuning of databases can be achieved.

In this paper, we deal with the functional and approximate dependency inference problem by pointing out some relationships between relational database theory and Formal Concept Analysis. More precisely, the notion of functional dependency in database is compared to the notion of implication in Formal Concept Analysis.

We propose a framework and several algorithms for mining these dependencies from large database relations. The common data centric step of this framework is the discovery of agree sets, which are closed sets with respect to the closure operator for functional dependency. Two approaches for discovering agree sets from database relations are proposed: the former is a database approach based on SQL queries and the latter is a data mining approach based on partitions. Experiments were performed in order to compare the two proposed methods.

Inclusion dependencies together with functional dependencies form the most important data dependencies used in practice. Inclusion dependencies are important for various database applications such as database design and maintenance, semantic query optimization and efficient view maintenance of data warehouse. Existing approaches for discovering inclusion dependencies consist in producing the whole set of inclusion dependencies holding in a database, leaving the task of selecting the interesting ones to an expert user.

In this paper we take another look at the problem of discovering inclusion dependencies. We exploit the logical navigation, inherently available in relational databases through workloads of SQL statements, as a guess to automatically find out only interesting inclusion dependencies. This assumption leads us to devise a tractable algorithm for discovering interesting inclusion dependencies. Within this framework, approximate dependencies, i.e. inclusion dependencies which almost hold, are also considered.

As an example, we present a novel application, namely self-tuning the logical database design, where the discovered inclusion dependencies can be used effectively.

Given a set F of functional dependencies (FDs), Armstrong relations for F are example relations satisfying exactly F. Instead of starting from F, an interesting issue is to consider an existing relation, say r, and compute Armstrong relations for dep(r), the set of FDs satisfied in r. In this setting, the main contribution of this paper is to define so ­called Infor­mative Armstrong Relations (IAR), say s, for r such that s is a subset of r and s is an Armstrong relation for dep(r). Such a relation always exists since r itself is obviously an IAR for dep(r), but the challenge is to compute IAR whose size is as small as possible. First, we proof that generating the smallest IAR is NP­complete. Then, we give an heuristic to construct small IARs for a given relation. Some expe­riments have been performed on relations available in the KDD archive; they point out the interest of IARs to sample existing relations.

Inclusion dependencies together with functional dependencies form the most fundamental data dependencies used in practice. They are respectively the generalization of foreign keys and keys. Their utility is important for all applications in which data semantic is important: For example to perform evolution or maintenance of existing databases, or to construct a data warehouse from production databases. In this paper we propose a levelwise algorithm to discover inclusion dependencies holding in a database. We use an existing framework, in which we have made the following contributions: an original algorithm to generate level i+1 candidates from level i IND, a coherent method to generate candidate INDs for level 1, an implementation of the proposed algorithm and experimental results on real-life database. Despite the inherent complexity of this problem, performance evaluations show the feasibility of our proposal.

Extracting information about data semantics from existing databases is essential in any re-engineering process. Several sources of information can be relevant for tackling this task, e.g. physical schema, database extension or application programs. In this paper, we present a tool called DBA Companion which can be an help to deal with the understanding of existing relational databases.

The tool is based on the notion of agree sets. Agree sets allow to devise a framework dealing with a wide range of design problems. Agree sets are seen as a common data centric step of several algorithms useful for database analysis. These algorithms are: functional and approximate dependency inference, minimal key inference, example relation generation and normal form tests.

We show how the framework is integrated in a tool which can provide some help to database administrators or analysts. We consider two problems: functional dependency inference and approximate functional dependency inference.

In this paper, we propose a framework for a broad class of data mining algorithms for understanding existing databases: Functional and approximate dependency inference, minimal key inference, example relation generation and normal form tests. We point out that the common data centric step of these algorithms is the discovery of agree sets.

A set-oriented approach for discovering agree sets from database relations based on SQL queries is proposed. Experiments have been performed in order to compare the proposed approach with a data mining approach. We present also a novel way to extract approximate functional dependencies having minimal errors from agree sets.

In this paper, we propose a new efficient algorithm called Dep-Miner for discovering minimal non-trivial functional dependencies from large databases. Based on theoretical foundations, our approach combines the discovery of functional dependencies along with the construction of real-world Armstrong relations (without additional execution time). These relations are small Armstrong relations taking their values in the initial relation. Discovering both minimal functional dependencies and real-world Armstrong relations facilitate the tasks of database administrators when maintaining and analyzing existing databases. We evaluate Dep-Miner performances by using a new benchmark database. Experimental results show both the efficiency of our approach compared to the best current algorithm (i.e. Tane), and the usefulness of real-world Armstrong relations.

In this paper, we propose a new efficient algorithm called Dep-Miner for discovering minimal non-trivial functional dependencies from large databases. Based on theoretical foundations, our approach combines the discovery of minimal functional dependencies along with the construction of real-world Armstrong relations (without additional execution time). These relations are small Armstrong relations taking their values in the initial relation. Discovering both minimal functional dependencies and real-world Armstrong relations facilitate the tasks of database administrators when maintaining and analyzing existing databases. We evaluate Dep-Miner performances by using a benchmark database. Experimental results show both the efficiency of our approach compared to the best current algorithm (i.e. Tane) described in a detailed way in this article, and the usefulness of real-world Armstrong relations.

In this paper, we define a framework in order to deal with a broad class of data mining algorithms for database relation analysis: Functional dependency inference, minimal key inference, sampling database relations and testing normal forms. We point out that the common data centric step of these algorithms is the discovery of agree sets. Within this framework, we give a new characterization of left-hand sides of minimal functional dependencies from which two levelwise algorithms are devised for computing functional dependencies and minimal keys.

We propose two approaches for discovering agree sets from database relations: The former is based on SQL queries while the latter makes use of a particular implementation technique based on stripped partitions. Experiments have been performed in order to compare the two approaches.

Discovering data dependencies consists in producing the whole set of a given class of data dependencies holding in a database, the task of selecting the interesting ones being usually left to an expert user. In this paper, we take another look at the problems of discovering inclusion and functional dependencies in relational databases. We define rigourously the so-called logical navigation from a workload of SQL statements. This assumption leads us to devise tractable algorithms for discovering interesting inclusion and functional dependencies.