create small cubes includes from a big one

  • Hi,

    I have a new problem:)

    I will create a new cube from my DWH. But I couldn't decide how to do.

    This one is a little different from the others. Normally for a special subject I've created a simple cube with its 13 dimensions and its measures(nearly 60-70 measures for each cube)

    This time my subject is small different from other subjects. It has tittle and subtitles.

    For exemple title is charry, and the sub titles are "global, bancaire, financial "I want to group the cube under these subtitles?

    I've heared a notion as sub-cube??But i want to add that for each syubtitle I hav edifferent measures to analyse But I have the sama dimension

    Is there anyone who can help me?

    Thanx

  • You could use CubeSlice (link) to create one (or many) local cubes, though it must have been a while since I visited as I don't remember the price being that high previously :crazy:

    You can also (reasonably easily) create local cubes using SSMS. Googling for 'create local cube' plus maybe adding analysis services and/or XMLA should result in a lot of hits.

    You can also do the same thing in Excel with little effort.

    And if you're not actually after local cubes but more a subset, you could simply add a new cube to your AS database and then use filters on the partition/s to not include specific transactional data.

    HTH,

    Steve.

  • Here's what I suggest. My understanding of your requirements might be slightly off but here goes -

    1. Include all the measures in one large fact table

    2. Create perspectives for each different 'flavor' of the cube you want to expose

    Something like this might be easier when the next version of SSAS comes out as well... shhh... 😉


    Cheers,

    Ben Sullins
    bensullins.com
    Beer is my primary key...

  • Thanx for your responses

    I thougt to put all the measures to a big fact table but in my repport I want to visulaise that in a big cube the user can choose a a smaller subject (that i imagine a subcube). In this way he/she can analyse the measures of that subcube?

    Is it clear?:(

  • Not sure it's super clear (to me anyways 😉 ) but what Ben suggested would let you create one larger cube (which you can make visible to your users, if you want) but also the perspectives will appear to be 'smaller' cubes (assuming you exclude dimensions and/or measures) which are also exposed to your users.

    Steve.

  • We have 1 large cube with everything, and off of that we have perspectives. This sounds like what you are looking for.

    http://msdn.microsoft.com/en-us/library/ms174803(SQL.90).aspx

    Greg E

  • it was exactly the respons that i was looking for. I applied it immedetly. Now my big cube can be selected in reporting serviceses as the partitions.

    I've decided to create a cube with different group of measures but

    as my dimensions are the same for all the group of measures. There were the joins that decrease the performance of the queries "select".

    This is a important point that we make atttention.

    Thank you all soo much for your responses

    merci

Viewing 7 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic. Login to reply