Dwarfs in the rearview mirror: how big are they really?
Proceedings of the VLDB Endowment
FCLOS: A client-server architecture for mobile OLAP
Data & Knowledge Engineering
Computing iceberg quotient cubes with bounding
DaWaK'06 Proceedings of the 8th international conference on Data Warehousing and Knowledge Discovery
Hi-index | 0.00 |
Iceberg Dwarf (IceDwarf for short) combines thestrength of Iceberg-Cube and Dwarf. It exploits the elegantDwarf structure for cube tuple store and eliminates thoseunsatisfied sub-dwarfs. By only storing nontrivial cubetuples, IceDwarf reduces the size of a dwarf significantly;even Dwarf itself compresses the data cube effectively.We studied how to efficiently compute icedwarfs, anddeveloped a straightforward algorithm (PAC). To furtherimprove the performance, we explored the structure ofDwarf and presented four nice lemmas. Based on theseobservations, we proposed a new algorithm called PWC. Itbuilds the IceDwarf by bottom-up computing all thepartitions of a fact table and inserting them into the Dwarfstructure, enabling Apriori-like pruning and single tuplepartition optimization, and facilitating the detection ofsuffix redundancies. Our performance study showed thatPWC is highly efficient and runs much faster than PAC foricedwarfs, even for computing full dwarfs.