Visual Construction Of Scientific Data Queries And Query Processing Optimization Techniques

Huge amounts of scientific data coming made unprecedented scientific data management difficulties faced by scientists in the field. On one hand, traditional database management system's table data model cannot meet the array data model needs of large-scale scientific data management. Besides, scientific data analysis tasks usually expressed in the form of analytical queries, for non-scientists in the field of computer science backgrounds, however, write complex queries using specific query language is still difficult.Therefore, to solve the difficult problem of building complex query statements to define scientific data analysis tasks for domain-specific scientists, and to optimize complex queries' processing is of great value and significance. This paper studies the demand of scientific data management and analysis, and then focuses on the graphic query building and performance optimization of complex queries' processing, combining with scientific data management and distributed query optimization technology. The main research contents of this paper include(1) a visualized analytical query building strategy is presented to lighten the scientists' overweight burden to learn complex syntax and semantic rules of complex query languages;(2) a scientific data analysis system named FASTDS is designed and implemented to enable analytic tasks can be composed by dragging and dropping, and currently, this system has been successfully applied in astronomic data;(3) to balance performance between data loading and query processing, a chunk size selection strategy is proposed;(4) the queries processing performance has been improved, respectively based on scheduling strategy with single-query optimizing, and using multi-queries optimizing.