This dataset is related to DRL-assisted Sensing-aware MAC scheduling. It contains temporal evolution of a communication scenario involving a single 5G gNB and multiple UEs running XR services. Over that communication scenario, sensing information becomes available to the RAN Intelligent controller. Then, such sensing information is used to adapt the MAC scheduling policy considering both QoS service requirements and radio resource utilisation. The dataset contains metrics used for MAC scheduling, including RLC buffer occupancies and MCS. Furthermore, traffic application metrics are also recorded, according to the kind of XR service. On top of that, sensing information used to optimize the MAC scheduling decisions are also kept, with the UE location and obstacle detection. The data has been generated using the ns-3 5G-LENA simulator, which implements the two-level MAC scheduling. An xApp gathers RAN metrics (including sensing data) and communicates with the MAC scheduler to support resource allocation decisions.