Search

In property graph query languages, nothing other than the properties of the current edge and those of its two vertex endpoints can be used to decide if an otherwise qualifying edge should be added to extend the current variable-length path. This, or the even stricter “thou can’t see anything” commandment in SPARQL arbitrary-length property path, is an impediment to writing queries that require that the inclusion (or not) of an edge be contingent upon properties of its neighborhood. An example of such a query requirement, supported in SQL, would be finding gender-alternating foaf:knows paths (in RDF data), or finding (in a property graph) money transfer trails carrying the same amount of money throughout. Proposed fully optional constructs for neighborhood-aware path traversal that eliminate these restrictions without requiring any changes to the underlying graph data model may facilitate expedited widespread adoption of graphs for representing enterprise data., This presentation discusses the motivation and current state of the art for Neighborhood-Aware Path Traversal (NAPT) in graph query languages. It proposes an approach that is rooted at introducing notations for references to neighborhood elements of the current edge being processed. Ideas are explained for new constructs that enable specification of constraints involving the properties of the neighborhood elements of an edge so that only those edges that satisfy the constraints are added to the path being created. Additional constructs are proposed for retrieval of path (aggregates), aggregate-based filter, and returning top-K paths based on specified criteria.

This presentation discusses the motivation and current state of the art for Neighborhood-Aware Path Traversal in graph query languages. It proposes an approach that is rooted at introducing notations for referencesto neighborhood elements of the current edge being processed. Ideas are explained for new constructs that enable specification of constraints involving the properties of the neighborhood elements of an edge so thatonly those edges that satisfy the constraints are added to thepath being created. Additional constructs are proposed for retrieval of path (aggregates), aggregate-based filter, and returning top-K paths based on specified criteria.

Data vocabularies facilitate the organization and retrieval of knowledge. As the volumes of data being generated in the internet age is exploding, the need for structuring data such that automated organization can occur is becoming even more crucial to manage this data. A popular method of searching this data in the context of a vocabulary is through the use of graph theory. This paper describes the scenario where, through the use of a data vocabulary for describing automotive data based on labeled property graphs, data is modified such that it is stored in a knowledge graph. The difficulties that were encountered in this effort and how they were overcome are also discussed.

Modeling your data as a graph has a significant advantage: The schema does not need to be explicitly defined or specified ahead of time. Thus, you can add data to your graph without being constrained by any schema. One of the less recognized problems with data addition to a graph, however, is the potential for loss of backward compatibility with regard to queries designed before the changes are made to the data. Use of RDF Quads (W3C RDF1.1 Recommendation 25-FEB-2014) as your graph data model would allow schema evolution caused by data addition to your graph to preserve backward compatibility of all your pre-existing queries. The tutorial also includes discussions on the following topics: new way of distinguishing among the various types of graphs: normal (discrete math) graph, RDF triples, Property Graph, RDF Quads. the new RDF# proposal for extension to RDF and SPARQL that allows explicit naming of RDF triples review of W3C SPARQL Query language and W3C SPARQL Update (used for DML) comparison between two popular graph query languages: SPARQL and PGQL use of RDF data for graph analytics review of W3C R2RML (RDB to RDF Mapping Language) methodology for creating relational-to-RDF (R2RML) mapping for a given relational schema. All of these discussions are complemented with live demo on a freely downloadable VM running Oracle 19c. &nbsp