getOrderList API should provide one unioned, sorted result set across active and history records

Today's implementation of getOrderList API makes two calls under the hood - one to archive tables and one to active tables (assuming archive switch was invoked in the call.)

Almost *no* consumer of order history wants two separate lists, each sorted with its own history. Rather, the consuming application typically wants one cohesive time-ordered list of orders, especially from an Order Management system of record. How the data is stored and retrieved in the underlying should be of no concern to the consuming application(s), and the ramifications of the underlying design should be masked for the consumer. The serving application should take the onus to union THEN sort the result set, rather than requiring comparable functions in each and every consuming system.

Presumably, this behavior was designed with an assumption that archive records are inherently older that active records. This is not always true. As a multi-brand retailer, with staggered onboarding of brands to OMS, it is entirely possible to have a single customer with migrated (archived) records which are newer than one or more of the active records received for another brand. The net result to the customer is a brand-diminishing, improperly sorted order history (Several customers have already called this out.)

There is also a behavior in Call Center which moves order records from archived to active upon search and view, even when no modifications are made, which further exacerbates the issue.