The percentage cube
作者:
Highlights:
• It is necessary to adapt SQL query processing to evaluate percentage cubes efficiently.
• A percentage cube is significantly more difficult to compute than a standard cube due to a higher exponential complexity.
• Percentage cubes should be computed at low cube dimensionality with dimensions having low cardinality.
• Selecting top-k percentages across all cuboids is the most difficult analysis, harder than selecting minimum percentages.
• Incremental materialized view algorithms are feasible for one percentage query, but not for the percentage cube.
摘要
•It is necessary to adapt SQL query processing to evaluate percentage cubes efficiently.•A percentage cube is significantly more difficult to compute than a standard cube due to a higher exponential complexity.•Percentage cubes should be computed at low cube dimensionality with dimensions having low cardinality.•Selecting top-k percentages across all cuboids is the most difficult analysis, harder than selecting minimum percentages.•Incremental materialized view algorithms are feasible for one percentage query, but not for the percentage cube.
论文关键词:
论文评审过程:Received 10 July 2017, Revised 16 December 2017, Accepted 22 January 2018, Available online 31 January 2018, Version of Record 5 November 2018.
论文官网地址:https://doi.org/10.1016/j.is.2018.01.005