Učitavanje video playera...
01:09:29
Domagoj Vrgoč: How to evaluate navigational queries over graphs?
One of the biggest differences in querying relational data and graph data is the ability of graph query languages to specify recursive patterns which explore parts of the graph whose size is not bounded in advance. The most common feature of such languages is exploring paths in the graph. However, path queries have traditionally been a problematic feature in terms of query evaluation, and most existing systems struggle to execute them efficiently. In this talk we will speculate that this might be due to the fact that a proper baseline for path queries was never explored in sufficient depth by the database community. To support our claim, we will show how to adapt classical graph search algorithms for path queries, and how their implementation performs on real world datasets whose size is measured in terabytes. Additionally, we will illustrate how such algorithms can trivially support query features sought after by the upcoming ISO standard for graph query languages, such as returning all matching paths.
