.png){kind=icon}
.png){kind=icon}
Return
Efficient Updates for Worst-Case Optimal Join Triple Stores
.png){kind=icon}
PRE
AI
DOI：10.1007/978-3-032-09527-5_29.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
It has been recently shown that worst-case optimal joins can significantly speed up query processing in RDF triple stores, especially in analytical workloads. However, this increase in query speed comes at the expense of updates being slow or not supported at all. We see this limited compatibility with updates as a key reason for the slow adoption of worst-case optimal joins in triple stores. In this paper, we address this challenge by presenting a fast, incremental insertion and deletion algorithm for the hypertrie, a worst-case optimal join data structure. This update algorithm can be used for offline bulk updates as well as online updates. Our evaluation on realistic update loads from DBpedia and scaling update sizes on Wikidata shows that the online performance of our algorithm is comparable to or better than that of traditional triple stores.
Keywords:
SPARQL update
worst-case optimal join
graph database
data structure
indexing