Accelerating Power Big Data Queries by Binary Partitioning and Workload-Driven Pre-joins

Querying power big data across multiple sections plays an important role in achieving efficient power data sharing and emergent accident handling. However, power big data often suffers from the issues of data sparsity and low query efficiency. In this paper, we propose a power big data query acceleration method based on binary partitioning and workload-driven pre-joins (BP-WPJ). First, the original power property tables can be vertically partitioned into multiple binary tables (BTs) for eliminating null-value elements, producing the statistics of data type frequency in BTs. Second, a workload-driven pre-join processing method (WPJ) is proposed to identify the join patterns of the query statement in terms of predicates and join relations. Third, we optimize the order of query triples by dividing the query statement into multiple query triples based on the bounded elements for accelerating power data query. Experiments were made on three datasets compared to three popular methods, and the results show that BP-WPJ is very competitive against the three baselines. It can significantly accelerate big data query and achieves relatively lower memory consumption, thereby meeting the practical needs of power big data queries.