I have a star schema with:
- a fact table with one column per dimension (around 10 to 15 dimensions) and one key figure. Volume would be between 50 million to 100 million records
- a table per dimension, each dimension has attributes and can have one or several hierarchies:
- 3 dimensions are flat (only one level)
- rest of dimensions have multiple levels (from 3 to 12) and can also have multiple parallel hierarchies (up to 5 per dimension). Most of the dimension are so called ragged hierarchies (where the leaf members are on different levels). Dimension members will range from 10 to about 1000 members per dimension.
Reporting requirements:
- aggregations on hierarchies
- selection/filtering on attributes
Bonus requirement:
- Possibility to add calculations (like the calculation of a KPI on all levels)
Would MonetDB be a good tool for that use case, I could not really find any information about hierarchical (parent/child) aggregations ?
MonetDB would definitely be a good candidate. As always, specific use cases call for a little experimentation. Since you have the schema (and probably the queries and data) it would be afternoon exercise.