All geek questions in one place

Better data structure for multidimensional data?

I would like to implement a simple mechanism for storing OLAP cubes in memory for reading and writing (writeback) - functionally similar to an SSAS cube with several dimensions, but one dimension and with only 1 type of aggregation (sum). As in the OLAP cube, each axis in multidimensional space can be a multilevel hierarchy.

Could the community provide me with some tips on which to look at data structures and related algorithms? I understand that I need something that can index data in many dimensions at the same time and store intermediate pre-calculated aggregation values.

I would prefer not to glue several nested cards together, but to implement something from scratch - the goal of the exercise is not just to realize this beast, but also to better understand multidimensional data structures and algorithms.

Just to clarify - I focus on the basic data structure of storing multidimensional hierarchical data for reading and writing. I am not trying to implement the MDX parser, make the cube permanent, etc.

+6
source share
1 answer

Look at the list of spatial indexes on Wikipedia , one of them, for example R-tree or kd tree, may be what you are looking for.

+6
source

Source: https://habr.com/ru/post/901657/

More articles:


All Articles