Global Dimensions

Note: The Mashup Dimensions feature, that was released in November 2023, will in many ways be able to replace Global Dimensions. Link to Mashup Dimensions article: Mashup Dimensions

Global Dimensions is a feature that makes it possible to link otherwise unrelated hierarchies. Doing so enables you to perform only one filtration in the criteria bar, this filtration will then take effect on all the hierarchies in the same Global Dimension.

You create an XML file in which you connect hierarchies in a "global dimension" element.


The hierarchies must have the same structure, i.e. same type, same amount of levels, and same content. Anexample to clarify this: if you have a time hierarchy of 'YMD' you cannot combine it with a hierarchy with a 'YQMD' level, they have to be the same. Please note:

  • To create a Global Dimension you must provide it with a unique id.
  • Hierarchies that should be linked must be included in the same "global dimension" element. You can link as many hierarchies as you wish.

There is no user interface for creating Global Dimensions, so you must create an XML file and create the "global dimension" elements you need. Please follow the steps below and take a close look at the example on the next page, this should enable you to create your own Global Dimensions.

  1. Create a folder for your XML file. This folder must be located in the 'Settings' folder on the TARGIT Server.
    Default path is: 'C:\ProgramData\TARGIT\ANTServer\Settings'.
    The folder must be named 'GlobalDimensions'.
  2. Create an XML file named 'GlobalDimensions.xml' and save it in the folder above.
    You are now ready to create Global Dimensions in the XML file.
  3. Once you have created your Global Dimensions definitions in the XML file and saved it, you may need to restart your TARGIT server.

Please note:

If your data source is an OLAP database with mulitiple cubes in the same database, it it NOT possible to utilize global dimensions. 
In those use cases, the concept of shared dimensions (which is an OLAP feature) should cover your needs to filter across cubes.

On the next page you will find an example of an XLM file with two global dimensions elements: 'time_hierarchy" and "time_month".

Use the example and the information below to create your own Global Dimensions.

Below is an example of an XML file with two global dimensions elements: 'time_hierarchy" and "time_month":

<?xml version="1.0" encoding="utf-8"?>
<globalDimension id="time_hierarchy">
<metadata database="DemoData" cube="Sales" dimension="Period" hierarchy="Period" />
<metadata database="CasualClothing" cube="Clothing" dimension="Posting Period" hierarchy="Hierarchy" />
<globalDimension id="time_month">
<metadata database="DemoData" cube="Sales" dimension="Period" hierarchy="Month" />
<metadata database="CasualClothing" cube="Clothing" dimension="Posting Period" hierarchy="Month" />

The global dimension with id="time_hierarchy" in the example above links the "Period" hierarchy (in the "Period " dimension in the Sales cube) with the "Hierarchy" hierarchy (in the "Posting Period" dimension in the "CasualClothing" cube).

Likewise, the "Month" hierarchies in the "Period" and "Posting Period" dimension are linked in the global dimension with id="time_month".


Note: Remember that the hierarchies must have the same structure and also note that the key-values (which are used for Global Dimensions might not be the same as the displayed values - see example below):


Keyvalues vs. Display values

In Global Dimensions the values that are mapped "behind the scene" is actually key values and not so called display values.

In this example I will try to map some weekdays from one cube against some weekdays from another cube and illustrate this point.

One of the weekdays is from TARGITs demodata Salescube.

If I locate the weekday in the source data panel and right click it - I get an offer to see information about it:

The explanation tells me that the weekdays I see (Monday, Tuesday, Wednesday, Thursday etc. has a key that is an integer (1,2,3, 4 etc):


This is important information - it means that when I map Time Weekday against another dimension, that dimension will have to have a similar key.

This attribute from my Data Discovery data source won't do - even though it's showing the same weekdays (display value), the keys are not the same as the previous Weekday from demodata:


In my Data Discovery data source I have another attribute called weekday, where the information shown looks like this:

This will do - even though the displayed values are different - the keys are exactly the same as my Weekday from Demodata.

My Global Dimensions document will look like this:


So remember to look out for key-values when using Global Dimensions.




Was this article helpful?
1 out of 1 found this helpful


  • Hi, 

    I created a hierarchy as stated i example above and it works. However can we do this for non hierarchies as well? So only for the dimension. A hierarchy is built up from dimensions, suggesting grouping levels. What if I only want to have a global dimension on 1 dimension or on both the hierarchy and the individual dimensions.


    Dimension: Company Dim

    Hierarchy: Company

    Level 1: Department

    Level 2: Employee

    I want a global dimension on the hierarchy Company but also on level 1 Department and Level 2 Employee each individually (so 3 in total).

    What is the code then? 


    <metadata database="DemoData" cube="Sales" dimension="Company Dim" hierarchy="Company" />


    <metadata database="DemoData" cube="Sales" dimension="Company Dim" hierarchy="" />


  • Hi Dirk

    There are no limitations to the new mash-up dimension in terms of platforms.

    If your dimensions from Tabular and OLAP has the same key you should be fine (and of course it's only OLAP that operates with keys).

    Mash-up dimensions have been out there since the the 2023 November-2 release.

  • Hi John,

    See my reply to the Topic post here: Global dimensions

    BR / Ole

  • I remember my colleague requested a feature that combines Tabular & OLAP based dimensions together as a Global Dimension.

    What is the status of this request or probably when was it released?

  • It would be nice to have another interface in order to work with the global dimensions, kind of the way we work with Data Discovery in form of a sort of "drag and drop" way of connecting the different dimensions across the various cubes.


Please sign in to leave a comment.