Closed klaudiusz223 closed 11 years ago
@klaudiusz223 Thanks for the reporting this. It results from the fact that child members of '[Geography].[Israel]' is not direct decendents of that level.
I'm investigating if such a cube configuration is legal or not.
It seems that it's legal, as mentioned here on 'ragged'dimensions :
Definitely it is legal.
You can check it in mondrian documentation - hideMemberIf attribute http://mondrian.pentaho.com/documentation/xml_schema.php#Level.
Additionally you can look at parent child hierarchies. Something is also wrong.
SELECT {[Measures].[Number of Employees]} ON COLUMNS, Hierarchize(Union(Union(Union({[Employees].[All Employees]}, [Employees].[All Employees].Children), [Employees].[Sheri Nowmer].Children), [Employees].[Sheri Nowmer].[Derrick Whelply].Children)) ON ROWS FROM [HR]
@klaudiusz223 Thanks very much for finding them. I've filed a separate issue for parent-child hierarchies here : #77 .
I think they are quite major limitations with the project, so I'm gonna deal with them immediately.
IMO you are absolutely right. Since jpivot isn't under active development it is quite big problem if someone needs open source tool that supports ragged nad parent child hierarchies. For exampele saiku has even bigger problems with such hierarchies.
What I found so far :
I suppose we can get around this problem with some ugly hacks. But before this, I'll check if it is indeed an intended behavior from Olap4J or Mondrian.
Unique name problem seems to be a Mondrian issue:
Result of following query is incorectly rendered if "show parent" option is set
SELECT NON EMPTY {[Measures].[Unit Sales]} ON COLUMNS, NON EMPTY Hierarchize(Union(Union(Union({[Geography].[Canada], [Geography].[Israel], [Geography].[Mexico], [Geography].[USA], [Geography].[Vatican]}, [Geography].[Israel].Children), [Geography].[Mexico].Children), [Geography].[USA].Children)) ON ROWS FROM [Sales Ragged]