Hi Sanjay,
If you understood the concepts, it will automatically answer your query. 
As you might have understood, the partitioning depends upon the data volume. If your table is expected to contain data which can be logically partitioned on basis like time dimension, Geographical dimension etc, then depending upon the query pattern you can opt to partition the data based on these dimensions.
If you have multiple nodes, then HANA automatically partitions the data and stores on multiple nodes, but an explicit partition always helps from the query performance perspective.
No of partitions depend upon the grouping of the partition criteria.
In short, there is no straight forward answer to your question, but the factors like data volume, what data is stored (possible partition criteria), Query pattern etc decide how the partitioning should be done.
Regards,
Ravi