Data Cubes
SQL Frames allows creating multi-dimensional data cubes. It supports Slicing, Dicing, Rollup, Drill Up/Down and Pivot operations of the Data Cube.
The advanced GROUP BY constructs ROLLUP and CUBE allow creating multiple levels of details at once. Hence, it is possible to create a single multi-dimensional cube that can then be sliced and diced as needed. There are multiple advantages with creating a data cube.
- They allow drilling up and down the hierarchies such as geography and time.
- Higher level values are computed from lower level details which are already aggregated and hence efficient to compute.
- Having multiple levels together allows complex visualizations such as pivot tables with sub-totals and static grid charts with multiple levels of details.
- Aggregation is relatively expensive and hence keeping pre-aggregated data allows projecting only the necessary slices for a given visual. The cost of aggregation is thus shared across several visuals.