For Pentaho Anaysis data source, there are 3 types of OLAP data source types [Pentaho Analysis, Pentaho Analysis De-normalized, and Pentaho Analysis (Legacy)].  The difference between them are how the column names are created from the MDX statement and how the data is organized.

Based on the following MDX query using SteelWheels:

select NON EMPTY Crossjoin({[Measures].[Quantity]}, Union({[Markets].[All Markets]}, [Markets].[All Markets].Children)) ON COLUMNS,
  NON EMPTY Hierarchize(Union({[Product].[All Products]}, [Product].[All Products].Children)) ON ROWS
from [SteelWheelsSales]

JPivot returns it like this:

For the Pentaho Analysis datasource aka Banded MDX.  The columns return like this: 
[Measure].[Quantity]/[Markets].[All Markets],
[Measure].[Quantity]/[Markets].[All Markets].[APAC], 
[Measure].[Quantity]/[Markets].[All Markets].[EMEA], 
[Measure].[Quantity]/[Markets].[All Markets].[Japan], 
[Measure].[Quantity]/[Markets].[All Markets].[NA].

For the Pentaho Analysis  (De-Normalize) datasource aka Denormalized MDX.  The columns return like this where it de-normalizes the crosstab data: 

For the Pentaho Analysis (legacy) datasource aka Legacy Banded MDX, this is the old way the columns returned prior to Pentaho Report Designer 3.5 and created so that reports migrated easily to 3.5 and beyond.  The columns return like this:  
Quantity/All Markets/Markets,