Query simplification by removing unwanted columns from M query will significant performance improvement.
I understand database optimizer will take care of it. But when using HANA as relational and leveraging HANA Views, optimizer will not remove unwanted columns. HANA will move the data from column engine to row engine where the unwanted columns removed.
This means, the damage is done already. Often moving large numbers of columns from column engine to row engine will consume significant amount of memory in HANA, resulting in unstable platform.
We overcome this by creating several M queries (to meet several scenarios) and do manage aggregate which is time consuming and complex.
By implementing query simplification, option can be provided in the query property to enable this feature, will generate optimized sql.