[Mondrian] RE: Same hierarchy on different axis

Julian Hyde jhyde at pentaho.com
Mon Jul 20 22:21:37 EDT 2009

Sorry folks, ain't gonna happen.
First off, Microsoft doesn't allow it in their MDX dialect. There's usually
a good reason for their language choices, such as...
Second, it doesn't make sense. If it's '1997.Q1' on the columns axis, and
'1997.Q2.4' on the rows axis, what value should the cell show? What context
should be provided when evaluating calculated members in that cell. Not well
Third, there's a workaround. Define and use attribute hierarchies. In the
above example, you probably wanted quarters on columns, and months on rows.
Do that by defining attribute hierarchies Time.Quarter and Time.Month in
addition to Time.Yearly and Time.Weekly.


From: mondrian-bounces at pentaho.org [mailto:mondrian-bounces at pentaho.org] On
Behalf Of Brian Vandenberg
Sent: Monday, July 20, 2009 12:00 PM
To: Mondrian developer mailing list
Subject: Re: [Mondrian] RE: Same hierarchy on different axis

  Ah, I should probably clarify there.  I skimmed what Julian wrote, sorry.
It would be nice if you could have the same hierarchy on multiple axes, I
was trying to express my enthusiasm toward the change Julian mentioned.

  If we could get both, all the better.


On Mon, Jul 20, 2009 at 12:52 PM, Pedro Alves <pmgalves at gmail.com> wrote:

Just to check - Brian, do you agree it does *not* make sense to allow the
same hierarchy in multiple axis? :( 

On 07/20/2009 03:37 PM, Brian Vandenberg wrote:

  I agree 100% on this.  Just being able to use multiple (unique)
hierarchies across multiple axes would be very convenient.


On Fri, Jul 17, 2009 at 12:37 PM, Julian Hyde <jhyde at pentaho.com

<mailto:jhyde at pentaho.com>> wrote:

   I don't think it makes sense to allow the same hierarchy on two
   different axes. What does make sense is what AS2005 does. They have
   'attribute hierarchies' which look a lot like levels except they can
   be used independently. But you still have old-fashioned hierarchies
   with multiple levels. And typically, an attribute that is used as a
   level would also have its own attribute hierarchy.
   E.g. the customer dimension might have a customer hierarchy, a state
   attribute hierarchy, a city attribute hierarchy, and a gender
   attribute hierarchy. You can have customer hierarchy on one axis,
   and state on another. But you can't have the customer hierarchy on
   two different axes.
   Hope that makes sense. If anyone has a good reference to attribute
   hierarchies, please share it. 'MDX Solutions', by Spofford et al.,
   explains it well, but I can't hyperlink a book sitting on my desk. :)
   PS Sorry to hear about your motherboard.

       *From:* Pedro Alves [mailto:pedro.alves at webdetails.pt
       <mailto:pedro.alves at webdetails.pt>]
       *Sent:* Friday, July 17, 2009 7:56 AM

       *To:* jhyde at pentaho.com <mailto:jhyde at pentaho.com> 

       *Subject:* Same hierarchy on different axis


       Sorry to reply to you directly; I was replying to the ML when my
       server's motherboard crashed, and now I have a nasty problem to
       solve. I will when I get it back, but wanted to reply to you
       since you told you might do some work on that issue

       You said that it will give error if members from the same
       hierarchy are used in different axis. Can this NOT happen?
       Better than explain, I'll show you why it's important for me:

       I attached 3 pictures of the designer I'm working in. I would
       like users to make the combinations they want. In that example,
       the user defined 2 selectors, one for the year and another for
       the market; Now he wants to do 2 charts - clients and countries
       (countries are in the same dimension as markets, just a
       different level).

       It would make my job much easier if I could just abstract what's
       displayed from what's being filtered; output.png shows the final
       result, of course, reflects a mondrian exception related to the
       same hierarchy being in 2 axis.

       IMHO it should be up to the users the responsibility to make
       sure he doesn't use weird combinations (like select
       [Countries].[Portugal] on 0, [Countries].[USA] on 1)



       Pedro Alvespedro.alves at webdetails.pt
<mailto:pedro.alves at webdetails.pt>
       WebDetails Consultingwww.webdetails.pt  <http://www.webdetails.pt>

   Mondrian mailing list
   Mondrian at pentaho.org <mailto:Mondrian at pentaho.org> 



Mondrian mailing list
Mondrian at pentaho.org

Mondrian mailing list
Mondrian at pentaho.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20090720/4db5e626/attachment.html 

More information about the Mondrian mailing list