The development of new types of power systems poses several key challenges for super-large-scale online analysis and optimization. These challenges include data merging, computation performance, and operation optimization. The uncertainty arising from renewable energy sources and load fluctuations has resulted in highly volatile power flow patterns compared to previous systems. As a result, the operating states of the system are more vulnerable to potential operational risks during critical periods. From the perspective of unit commitment, optimizing system resources involves increasingly complex constraints and variables. Graph computing has shown exceptional performance in online power system analysis, particularly in topology-related analysis. This paper explores the application of graph computing techniques to enhance security-constrained unit commitment historical data management. It offers new insights into the utilization of graph computing to improve unit commitment.