A Two-Stage Locating Supported Primary Key Indexing Method for Partitioned Real-Time Databases With Steady-State Timestamps

This paper proposes a partitioned-table primary key indexing method with secondary positioning capability for steady-state timestamped real-time databases in power automation systems. Addressing the performance degradation of traditional binary search when handling million-scale datasets in smart grid dispatching automation systems, the method incorporates bit-field partitioning, zone statistics, range positioning, and skip search techniques to achieve efficient subsecond-response data queries. Experimental results demonstrate that compared with conventional binary search, the proposed method achieves $\mathrm{O}(1)$ time complexity in continuous data distribution scenarios and reduces query iterations by approximately 30 % in discontinuous distribution scenarios, significantly enhancing the query performance of steady-state timestamped real-time databases.