These guidelines apply to standard and measure hierarchies:
When working with relational data sources:
Columns in a hierarchy can be added from columns in the Metadata Navigator or the Source Navigator.
The columns in a single-chain hierarchy must come from one or more tables within a single data source.
In a multichain hierarchy, each chain in the hierarchy may come from different data sources; however, within a single chain in the multichain hierarchy, columns must come from one or more tables within a single data source.
When creating a cube schema using a multichain hierarchy built from multiple data sources, you must override the data load binding on the lowest level (Gen1 or leaf level) of each chain in the hierarchy, as described in Defining Data Load Mappings.
When working with text file data sources, columns in a hierarchy—single-chain or multichain—must come from the same text file within a single data source.