Search

International audience; The standard notion of query answering over incomplete database is that of certain answers, guaranteeing correctness regardless of how incomplete data is interpreted. In majority of real-life databases,relations have numerical columns and queries use arithmetic and comparisons. Even though the notion of certain answers still applies,we explain that it becomes much more problematic in situations when missing data occurs in numerical columns. We propose a new general framework that allows us to assign a measure of certainty to query answers. We test it in the agnostic scenario where we do not have prior information about values of numerical attributes, similarly to the predominant approach in handling incomplete data which assumes that each null can be interpreted as an arbitrary value of the domain. The key technical challenge is the lack of a uniform distribution over the entire domain of numerical attributes, such as real numbers. We overcome this by associating the measure of certainty with the asymptotic behaviorof volumes of some subsets of the Euclidean space. We show that this measure is well-defined, and describe approaches to computing and approximating it. While it can be computationally hard, or result in an irrational number, even for simple constraints, we produce polynomial-time randomized approximation schemes with multiplicative guarantees for conjunctive queries, and with additive guarantees for arbitrary first-order queries. We also describe a set of experimental results to confirm the feasibility of this approach.

International audience; The Cypher property graph query language is an evolving language, originally designed and implemented as part of the Neo4j graph database, and it is currently used by several commercial database products and researchers. We describe Cypher 9, which is the first version of the language governed by the openCypher Implementers Group. We first introduce the language by example, and describe its uses in industry. We then provide a formal semantic definition of the core read-query features of Cypher, including its variant of the property graph data model, and its " ASCII Art " graph pattern matching mechanism for expressing subgraphs of interest to an application. We compare the features of Cypher to other property graph query languages, and describe extensions, at an advanced stage of development, which will form part of Cypher 10, turning the language into a compositional language which supports graph projections and multiple named graphs.

Query answering over incomplete data invariably relies on the standard notion of certain answers which gives a very coarse classification of query answers into those that are certain and those that are not. Here we propose to refine it by measuring how close an answer is certainty. This measure is defined as the probability that the query is true under a random interpretation of missing information in a database. Since there are infinitely many such interpretations, to pick one at random we adopt the approach used in the study of asymptotic properties and 0-1 laws for logical sentences, and define the measure as the limit of a sequence. We show that in the standard model of missing data, the 0-1 law is observed: this limit always exists and can be only 0 or 1 for a very large class of queries. Thus, query answers are either almost certainly true, or almost certainly false. We prove that almost certainly true answers are precisely those returned by the naive evaluation of the query. When databases satisfy constraints, the measure is defined as the conditional probability of the query being true if the constraints are true. This too is defined as a limit, and we prove that it always exists, can be an arbitrary rational number, and is computable. For some constraints, such as functional dependencies, the 0-1 law continues to hold. As another refinement of the notion of certainty, we introduce a comparison of query answers: an answer with a larger set of interpretations that make it true is better. We identify the precise complexity of such comparisons, and of finding sets of best answers, for first-order queries.

Data wrangling, the multi-faceted process by which the data required by an application is identified, extracted, cleaned and integrated, is often cumbersome and labor intensive. In this paper, we present an architecture that supports a complete data wrangling lifecycle, orchestrates components dynamically, builds on automation wherever possible, is informed by whatever data is available, refines automatically produced results in the light of feedback, takes into account the user’s priorities, and supports data scientists with diverse skill sets. The architecture is demonstrated in practice for wrangling property sales and open government data.

International audience; Schema mappings are a fundamental concept in data integration and exchange, and they have been thoroughly studied in different data models. For graph data, however, map-pings have been studied in a restricted context that, unlike real-life graph databases, completely disregards the data they store. Our main goal is to understand query answering under graph schema mappings – in particular, in exchange and integration of graph data – for graph databases that mix graph structure with data. We show that adding data querying alters the picture in a significant way. As the model, we use data graphs: a theoretical abstraction of property graphs employed by graph database implementations. We start by showing a very strong negative result: using the simplest form of nontrivial navigation in map-pings makes answering even simple queries that mix navigation and data undecidable. This result suggests that for the purposes of integration and exchange, schema mappings ought to exclude recursively defined navigation over target data. For such mappings and analogs of regular path queries that take data into account, query answering becomes de-cidable, although intractable. To restore tractability without imposing further restrictions on queries, we propose a new approach based on the use of null values that resemble usual nulls of relational DBMSs, as opposed to marked nulls one typically uses in integration and exchange tasks. If one moves away from path queries and considers more complex patterns, query answering becomes undecidable again, even for the simplest possible mappings.

XPath plays a prominent role as an XML navigational language due to several factors, including its ability to express queries of interest, its close connection to yardstick database query languages (e.g., first-order logic), and the low complexity of query evaluation for many fragments. Another common database model---graph databases---also requires a heavy use of navigation in queries; yet it largely adopts a different approach to querying, relying on reachability patterns expressed with regular constraints.Our goal here is to investigate the behavior and applicability of XPath-like languages for querying graph databases, concentrating on their expressiveness and complexity of query evaluation. We are particularly interested in a model of graph data that combines navigation through graphs with querying data held in the nodes, such as, for example, in a social network scenario. As navigational languages, we use analogs of core and regular XPath and augment them with various tests on data values. We relate these languages to first-order logic, its transitive closure extensions, and finite-variable fragments thereof, proving several capture results. In addition, we describe their relative expressive power. We then show that they behave very well computationally: they have a low-degree polynomial combined complexity, which becomes linear for several fragments. Furthermore, we introduce new types of tests for XPath languages that let them capture first-order logic with data comparisons and prove that the low complexity bounds continue to apply to such extended languages. Therefore, XPath-like languages seem to be very well-suited to query graphs.

The paper introduces and investigates the aggregation problem for synthesized mediators of Web services (SWMs). An SWM is a deterministic finite-state transducer defined in terms of templates for component services. Upon receiving an artifact, an SWM selects a set of available services from a library to realize its templates, and invokes those services to operate on the artifact, in parallel; it produces a numeric value as output (e.g., the total price of a package) by applying synthesis rules. Given an SWM, a library and an input artifact, the aggregation problem is to find a mapping from the component templates of the SWM to available services in the library that maximizes (or minimizes) the output. As opposed to the composition syntheses of Web services, the aggregation problem aims to optimize the realization of a given mediator, to best serve the [email protected]? need. We analyze this problem, and show that its complexity depends on the underlying graph of the mediator: while it is undecidable when such graphs contain even very simple cycles, it is solvable in single-exponential time in the size of the specification (i.e., the total size of the input SWM, library and artifact) for SWMs whose underlying graphs are acyclic. We prove several results of this kind, with matching lower bounds (NP and PSPACE), and analyze restrictions that lead to polynomial-time solutions. In addition, we study the aggregation problem for nondeterministic SWMs (NSWMs). We show that the aggregation problem for NSWMs with various underlying graphs retains the same complexity as its deterministic counterparts. We also provide complexity bounds for determining whether SWMs and NSWMs terminate.

Relational schema mappings have been extensively studied in connection with data integration and exchange problems, but mappings between XML schemas have not received the same amount of attention. Our goal is to develop a theory of expressive XML schema mappings. Such mappings should be able to use various forms of navigation in a document, and specify conditions on data values. We develop a language for XML schema mappings, and concentrate on three types of problems: static analysis of mappings, their complexity, and their composition. We look at static analysis problems related to various flavors of consistency: for example, whether it is possible to map some document of a source schema into a document of the target schema, or whether all documents of a source schema can be mapped. We classify the complexity of these problems. We then move to the complexity of mappings themselves, i.e., recognizing pairs of documents such that one can be mapped into the other, and provide a classification based on sets of features used in mappings. Finally we look at composition of XML schema mappings. We study its complexity and show that it is harder to achieve closure under composition for XML than for relational mappings. Nevertheless, we find a robust class of XML schema mappings that have good complexity properties and are closed under composition.

For many years, finite model theory was viewed as the backbone of database theory, and database theory in turn supplied finite model theory with key motivations and problems. By now, finite model theory has built a large arsenal of tools that can easily be used by database theoreticians without going to the basics such as combinatorial games. We survey such tools here, focusing not on how they are proved, but rather on how to apply them, as-is, in various questions that come up in database theory.

Design principles for XML schemas that eliminate redundancies and avoid update anomalies have been studied recently. Several normal forms, generalizing those for relational databases, have been proposed. All of them, however, are based on the assumption of anative XML storage, while in practice most of XML data is stored inrelational databases. In this paper we study XML design and normalization for relational storage of XML documents. To be able to relate and compare XML and relational designs, we use an information-theoretic framework that measures information content in relations and documents, with higher values corresponding to lower levels of redundancy. We show that most common relational storage schemes preserve the notion of being well-designed (i.e., anomalies- and redundancy-free). Thus,existing XML normal forms guarantee well-designed relational storagesas well. We further show that if this perfect option is not achievable, then a slight restriction on XML constraints guarantees a "second-best" relational design, according to possible values of the information-theoretic measure. We finally consider an edge-based relational representation of XML documents, and show that while it has similar information-theoretic properties with other relational representations, it can behave significantly worse in terms of enforcing integrity constraints.

A recently introduced information-theoretic approach to analyzing redundancies in database design was used to justify normal forms like BCNF that completely eliminate redundancies. The main notion is that of an information content of each datum in an instance (which is a number in [0,1]): the closer to 1, the less redundancy it carries. In practice, however, one usually settles for 3NF which, unlike BCNF, may not eliminate all redundancies but always guarantees dependency preservation.In this paper we use the information-theoretic approach to prove that 3NF is the best normal form if one needs to achieve dependency preservation. For each dependency-preserving normal form, we define the price of dependency preservation as an information-theoretic measure of redundancy that gets introduced to compensate for dependency preservation. This is a number in the [0,1] range: the smaller it is, the less redundancy a normal form guarantees. We prove that for every dependency-preserving normal form, the price of dependency preservation is at least 1/2, and it is precisely 1/2 for 3NF. Hence, 3NF has the least amount of redundancy among all dependency-preserving normal forms. We also show that, information-theoretically, unnormalized schemas have at least twice the amount of redundancy than schemas in 3NF.

Data exchange is the problem of finding an instance of a target schema, given an instance of a source schema and a specification of the relationship between the source and the target, and answering queries over target instances in a way that is semantically consistent with the information in the source. Theoretical foundations of data exchange have been actively explored recently. It was also noticed that the standard certain answers semantics may behave in very odd ways.In this paper I explain that this behavior is due to the fact that the presence of incomplete information in target instances has been ignored; in particular, proper query evaluation techniques for databases with nulls have not been used, and the distinction between closed and open world semantics has not been made. I present a concept of target solutions based on the closed world assumption, and show that the space of all solutions has two extreme points: the canonical universal solution and the core, well studied in data exchange. I show how to define semantics of query answering taking into account incomplete information, and show that the well-known anomalies go away with the new semantics. The paper also contains results on the complexity of query answering, upper approximations to queries (maybe-answers), and various extensions.

Data exchange is the problem of taking data structured under a source schema and creating an instance of a target schema. Given a source instance, there may be many solutions - target instances that satisfy the constraints of the data exchange problem. Previous work has identified two classes of desirable solutions: canonical universal solutions, and their cores. Query answering in data exchange amounts to rewriting a query over the target schema to another query that, over a materialized target instance, gives the result that is semantically consistent with the source. A basic question is then whether there exists a transformation sending a source instance into a solution over which target queries can be answered.We show that the answer is negative for many data exchange transformations that have structural properties similar to canonical universal solutions and cores. Namely, we prove that many such transformations preserve the local structure of the data. Using this notion, we further show that every target query rewritable over such a transformation cannot distinguish tuples whose neighborhoods in the source are similar. This gives us a first tool that helps check whether a query is rewritable, We also show that these results are robust: they hold for an extension of relational calculus with grouping and aggregates, and for two different semantics of query answering.

Normalization as a way of producing good relational database designs is a well-understood topic. However, the same problem of distinguishing well-designed databases from poorly designed ones arises in other data models, in particular, XML. While, in the relational world, the criteria for being well designed are usually very intuitive and clear to state, they become more obscure when one moves to more complex data models.Our goal is to provide a set of tools for testing when a condition on a database design, specified by a normal form , corresponds to a good design. We use techniques of information theory, and define a measure of information content of elements in a database with respect to a set of constraints. We first test this measure in the relational context, providing information-theoretic justification for familiar normal forms such as BCNF, 4NF, PJ/NF, 5NFR, DK/NF. We then show that the same measure applies in the XML context, which gives us a characterization of a recently introduced XML normal form called XNF. Finally, we look at information-theoretic criteria for justifying normalization algorithms.

XML specifications often consist of a type definition (typically, a DTD) and a set of integrity constraints. It has been shown previously that such specifications can be inconsistent, and thus it is often desirable to check consistency at compile-time. It is known that for general keys and foreign keys, and DTDs, the consistency problem is undecidable; however, it becomes NP-complete when all keys are one-attribute (unary), and tractable, if no foreign keys are used.In this paper, we consider a variety of constraints for XML data, and study the complexity of the consistency problem. Our main conclusion is that in the presence of foreign keys, compile-time verification of consistency is usually infeasible. We look at two types of constraints: absolute (that hold in the entire document), and relative (that only hold in a part of the document). For absolute constraints, we extend earlier decidability results to the case of multi-attribute keys and unary foreign keys, and to the case of constraints involving regular expressions, providing lower and upper bounds in both cases. For relative constraints, we show that even for unary constraints, the consistency problem is undecidable. We also establish a number of restricted decidable cases.