Abstract:Based on Hadoop framework"s distributed storage, parallel computing and data warehouse modeling techniques, this paper constructs Kylin multi-dimensional analysis system, which realizes the unified storage, calculation and analysis functions of oil and gas field injection well production data. It solves the problems of low efficiency of multi-table joint query and data islands between multiple databases. The fast query and management optimization of 16 distributed databases has been realized, and the efficiency of query analysis has been improved by 3 times. The platform can uniformly manage the production data of 8×104 injection wells. The business analysis time is shortened from 1 d to 5 s, and the query time is second response. Through the establishment of multi-dimensional analysis model of injection well production data, the application of macroscopic management analysis, problem well management analysis and injection well production operation analysis of injection wells was realized in China National Petroleum Corporation. The system responds quickly and meets the needs of efficient analysis. The injection molding production data analysis granularity is refined from the original oilfield to a single well, and the business analysis is more detailed, and the oil and gas production dynamics can be grasped in real time.