I have a relatively wide table (50 columns). Of those, the first 5 represent the scope of a certain subset of data and the last 45 represent properties (mapped at deployment time, so static). The users have the possibility (through a query builder) to query against such table and any combination of its property columns.
I am now wondering what would be the best strategy to build indexes for this table. Performance-wise, would it be better to create a large number of "small" indexes or a small number of indexes spanning many columns?