A Graph Query Processor for Queries of Class CRPQagg

在这个项目中，我们将通过设计和开发一个通用的图形数据库查询处理器来解决传统数据库研究中的这些需求。在这个项目中进行的研究将分为两个工作流。第一个工作流将从属于一个定义良好的查询语言类的具体图查询语言开始，“自上而下”地处理问题。本课程将分析流行的查询语言的功能，以便正式定义一个通用的图数据模型和统一的运算符代数。第二个工作流将通过从现有的图操作和索引结构的算法开始，“自底向上”地解决这个问题。将系统地研究图的特征与完成某种处理任务的可能算法之间的依赖关系。基于这一实证研究，将得出一个分析成本模型，将两个工作流的结果联系在一起，以建立一个查询处理器，将逻辑图形处理任务转化为优化的物理执行计划。

In this project, we will address these requirements in the tradition of database research by designing and developing a general query processor for graph databases. The research conducted in this project will be structured into two streams of work. The first stream of work will approach the problem "top-down" by starting from concrete graph query languages that belong to a well-defined class of query languages. The functionality of popular query languages in this class will be analyzed in order to formally define a common graph data model and a unified algebra of operators. The second stream of work will address the problem "bottom-up" by starting from existing algorithms for graph operations and index structures. The dependencies between the characteristics of a graph and the possible algorithms to accomplish a certain processing task will be studied systematically. Based on this empirical study, an analytical cost model will be derived that will tie the results of both streams of work together in order to build a query processor that translates logical graph processing tasks into optimized physical execution plans.