Let's say a data warehouse view has joins between a fact sales table, dim customer, dim location, dim product, and dim date. If all columns from all 5 tables were brought into the view, the number of columns could reach 200+. (keys are indexed)
If a query was written against this view that only brought in sales by price, customer name, location name, product name and datetime and didn't include any of the other columns available in the view, would it perform faster than a view that only brought in price, customer name, location name, product name and datetime?
Please contrast how does SQL Server handle queries against a view compared to a query against the same tables in that view when the query has less columns than are available in the view?
We are currently using SQL Server 2012.
