Selection steps differ fundamentally from conventional filters in their position within analytical processing. Oracle states that filters are evaluated before aggregation , whereas selection steps are applied after the query has been aggregated . As a result, selection steps determine which members are displayed without recalculating the previously established aggregate values.

For example, assume a hierarchy contains members totaling 100. If an ordinary filter removes one member before aggregation, the measure total can change. If a selection step removes that member after aggregation, the displayed membership changes while the aggregate can remain 100. This behavior is precisely why selection steps are valuable for analytical member manipulation such as Keep Only, Add, Remove, or condition-based selections.

Within classic analyses, Oracle supports creation of selection steps for attribute columns and hierarchical columns . Measure columns can be used to formulate conditions for selection steps, but selection steps aren't directly created for measures themselves.

Selection steps don't replace filters; both serve distinct analytical purposes. They also aren't restricted to pivot tables.

Therefore, C and D correctly identify their applicable column types and their post-aggregation effect on displayed members.

Reference Topic: Formatting Data and Managing Filters, Selection Steps, and Variables — Selection Steps