Section: New Results

Querying Heterogeneous Linked Data

Participants : Guillaume Bagan, Iovka Boneva, Angela Bonifati, Pierre Bourhis, Radu Ciucanu, Tom Sebastian, Slawomir Staworko, Sophie Tison.

Staworko, Ciucanu and Boneva presented a new class of schemas for unordered XML trees, which are based on unordered regular expressions, also called multiplicity schemas. They show that many static analysis problems become feasible when removing disjunctions there [6] .

Ciucanu and Staworko [8] investigated the case of unordered XML, where the relative order among siblings is ignored, and focused on the problem of learning schemas from examples given by the user. They considered disjunctive multiplicity schemas (DMS) and their restrictions, disjunction-free multiplicity schemas (MS). For both DMS and MS, they prove the learnable cases.

Regular path queries in graphs have found much recent interest in the context of Sparql queries for linked open data in the RDF format. Bagan, Bonifati and Groz (former PhD student of Mostrare, now PostDoc at Tel-Aviv University) have obtained a precise characterization of those regular path queries that can be answered with polynomial data complexity [5] leading to a trichotomy (AC0, NL-complete, or else NP-complete). Thereby, they have solved an open question (raised by W. Martens in PODS'12).

XPath query evaluation over compressed trees has been studied in [12] . They focused on a fragment of XPath, which is the downward, navigational XPath and presented precise bounds on the time complexity of XPath query execution over grammar-compressed trees. In particular, they focused on counting the nodes selected by an XPath expression, extracting and materializing their pre-order numbers and serializing the obtained subtrees.

In [2] , Groz, Staworko, Caron, Roos and Tison studied query rewriting with views when the classes used to define queries and views are Regular XPath and MSO. Next, they investigated problems of static analysis of security access specifications (SAS) by introducing the novel class of interval-bounded SAS and they defined three different manners to compare views (i.e. queries), with a security point of view. Finally, they provided a systematic study of the complexity for deciding these three comparisons.