Enhancement of Data Aggregation Using A Novel Point Access Method

The B + -tree and its variants have been reported as the good index structures for retrieving data. Database systems frequently establish the B + -tree style indices for fast access to data records. However, traditional B + -tree index could be a performance bottleneck because of its inflatable hierarchy. Many works focus on improving indexing techniques. In fact, the optimization of data organization inside index nodes is the most critical factor to improve retrieval quality. Some handles like pre-partition of data space, node splitting by force, node splitting with unbalanced partition, and node splitting upon overflow loading always burden index structures with plenty of storage space and building overhead. In this paper, we propose a new index scheme to highly aggregate the external structure in a B + -tree. It also adopts a better splitting policy to completely remove the suffering from data insertion orders. Our new index technique can compress data records in leaves and in turn reduce index size to improve query performance. In addition, the entire index's space utilization is promoted to a higher level; thereby the index's space requirement become s smaller and easily resides in memory.