How should 'time to access' be calculated for a dataset with multiple components that have different access times (e.g., a mix of instantly available open data and restricted data that requires approval)?

How should 'time to access' be calculated for a dataset with multiple components that have different access times (e.g., a mix of instantly available open data and restricted data that requires approval)?

For a complex dataset with components that have different access times, the reported "time to access" should always be the longest time required. In other words, the metric must represent the total time until all requested components of the dataset are fully available to the user.