According to our research, we have found several inconsistencies between the way the MS OLAP provider handles result cube data within an ADODB recordset and an ADOMD cellset. For example the following MDX query generates slightly different data using ADODB versus ADOMD:

MEMBER [Marital Status].[All Members] AS' aggregate({[Marital Status].members})'
MEMBER [Gender].[All Members] AS' aggregate({[Gender].members})'
MEMBER [Store Type].[All Members] AS' aggregate({[Store Type].members})'
MEMBER [Time].[Q3&Q4] AS' aggregate({[Time].[1997].[Q3],[Time].
MEMBER [Measures].[Unit Sales 1] AS' ([Measures].[Unit Sales])'
({[Gender].allmembers} * {[Measures].[Unit Sales 1]} ) ON
({[Store Type].allmembers} * {[Time].allmembers} * {[Marital
Status].allmembers}) ON ROWS
FROM [Sales]

When a user runs the above MDX query using ADOMD (on the MDX Sample Application), the user gets all members returned in the result cube from the following dimensions on the row axis: Store Type, Time, and Marital Status.

Yet, when the same query is run using ADODB, the user fails to get any
metadata for the "ALL" level (there are simply blank spaces on the
ensuing levels) on the row axis. In the above query this includes the
following "ALL" level members:

[Marital Status].[All Marital Status] - Real Member
[Marital Status].[All Members] - Calculated Member
[Store Type].[All Store Type] - Real Member
[Store Type].[All Members] - Calculated Member

In addition, it appears that ADODB DOES NOT support any format
properties in the recordset.

Is there a solution to the above problem? Are we overlooking any
other significant MDX limitations related to ADODB recordsets? Most importantly, does anybody know whether Microsoft is committed to fixing these errors?



Raj C.