This paper studies a novel form of last-mile delivery that combines delivery optimization of E-commerce and Online-to-Offline (O2O) parcels. While O2O parcels must first be picked up from O2O shops before being delivered to the appropriate customers, E-commerce parcels are delivered to customers from a predetermined number of depots. The goal is to use the same fleet of vehicles to provide integrated delivery services for various merchant types. We propose a branch-and-price algorithm based on the set partitioning formulation, where the pricing problem is solved by a label-setting algorithm. Several acceleration techniques including elementary path relaxation, ng-path relaxation and decremental space search are designed to speed up the price problem. Both the conventional dataset used in the literature and a real-world database with up to 75 clients are utilized to test the performance of the suggested method, demonstrating its efficacy.