Just to clarify your approaches, comments in <brackets>
1) Use the Attribute views and perform the joins as mentioned above.
<this is the standard HANA approach for analytic views>
2) Don't use the attribute views, but perform the joins in the Data foundation itself. In this approach, you can join the data foundation "fact table" with the attribute view tables (attribute and that to the text table).
<I assume you don't mean "don't use the attribute view", rather "use the natural key from the fact RATHER than the attribute view", correct?>
The Front end tool used is a third party tool which interfaces with HANA over JDBC / ODBC interface with SQL statements fired against the HANA Information models.
<So when you access a given model from the tool, what MATNR would be presented? Both or only one? If only one, which one is shown? Or does your tool only fire SQL statements without any 'native' UI?>
Thanks,
Justin