Abstract | ||
---|---|---|
A data warehouse is a collection of data from different sources that supports analytical querying. A Bitmap Index (BI) allows fast access to individual attribute values that are needed to answer a query by representing the values of an attribute for all tuples separately, as bit strings. A Property Map (PMap) is a multidimensional indexing technique that pre-computes attribute expressions, called properties, for each tuple and stores the results as bit strings [DD97, LD02]. This paper compares the performance of the PMap and the Range-Encoded Bit-Sliced Index (REBSI) [CI98] using cost models to simulate their storage and query processing costs for different kinds of queries over a benchmark schema. We identify parameters that affect performance of these indexes and determine situations in which either technique gives significant improvement over the other. We also explore ways to improve PMap design to enhance performance. |
Year | DOI | Venue |
---|---|---|
2002 | 10.1145/583890.583900 | DOLAP |
Keywords | Field | DocType |
range-encoded bit-sliced index,different source,multidimensional indexing technique,performance comparison,property map,individual attribute value,query processing cost,pmap design,data warehouse,bit string,different kind,bitmap index,bitmap indexing,indexation | Data warehouse,Data mining,Bitmap index,Information retrieval,Expression (mathematics),Bitmap indexing,Tuple,Computer science,Multidimensional indexing,Schema (psychology),Database | Conference |
ISBN | Citations | PageRank |
1-58113-590-4 | 3 | 0.49 |
References | Authors | |
5 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ashima Gupta | 1 | 93 | 9.17 |
Karen C. Davis | 2 | 247 | 39.78 |
Jennifer Grommon-Litton | 3 | 5 | 0.90 |