Hitachi Vantara Pentaho Community Forums
Results 1 to 2 of 2

Thread: One cube versus many cubes?

  1. #1
    Join Date
    Oct 2012

    Question One cube versus many cubes?

    I am puzzled about the correct way to design a cube vs. several cubes so that reports and queries that compare several different measurements can be made as easily as possible. The problem is that when there are different measurements, these measurements may only share some dimensions or some levels of a hierarchical dimension.
    For example (a very artificial example just to illustrate the question) a measurement about working hours may be available on a daily basis, involve the dimension male/female, job category, and also the dimension overtime. Another measurement e.g. production may be available only on a monthly basis, and have the dimension product type. Both measurements share the hierarchies year and month of the time dimension.
    These measurements are very different and only share a subset of dimensions, and dimension values, yet we want to create reports and queries where one measurement is compared with the other. For example we want to compare the total female working hours of job category A with the total production of product type X on a monthly basis. Or we want to compare the average daily working hours of all males with total production on a monthly basis.
    The usual advice given is not to make the cube to big, avoid "unrelated" measures and avoid too many dimensions, but for some applications, reporting over many measures that are related only by e.g. the time dimension is important. There seems to be no easy, "standard" way to create reports over many cubes in an ad-hoc fashion.

    The real problem we are dealing with actually requires more than two measurements: there are many measurements and there are some dimensions that are shared by all those measurements. There are other dimensions that are shared by just some measurements and there are still other dimensions that are unique to individual measurements. Practically all reporting/analysis tasks need to compare two or more of these measurements in many different ways. We could implement this in pure SQL manually but that would also mean to implement a lot of the query and web UI functionality etc. So the question is - how can this be accomplished, if at all, with standard OLAP and reporting tools and what modeling considerations does one have to keep in mind?

    So - how is this kind of problem usually dealt with? Is there some good reading stuff (wiki articles, books, etc.) for this?

  2. #2
    Join Date
    Oct 2012


    If this the wrong forum to ask this question?
    Is this a silly question to ask in the first place or does nobody know the answer? May there be something unclear in my question that would have to get clarified?
    Last edited by josmi; 11-09-2012 at 10:08 AM.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
Privacy Policy | Legal Notices | Safe Harbor Privacy Policy

Copyright © 2005 - 2019 Hitachi Vantara Corporation. All Rights Reserved.