Computing closest common subexpressions for view selection problems

Abstract

Selecting a set of views for materialization is a required task in many current database and data warehousing applications including the design of a data warehouse, and the maintenance of multiple materialized views. The selected views can be materialized permanently or transiently depending on the specific view selection problem. The view selection algorithms are expensive due to the size of the search space of the problem.In this paper we propose an approach for generating candidate views for materialization for view selection problems based on the definition of the input queries. We also provide rewritings of the input queries using the generated candidate views. In generating candidate views, we do not apply costbased techniques but we try to maximize the operations in the views. Subsequently, view selection algorithms can exploit problem dependent cost functions to choose among the generated candidate views. Our approach is not restricted to a specific view selection problem. Compared to a previous one, it generates views that involve more relation occurrences (or operations) and can reduce the size of the search space which can be very large. We implement our approach and we report some experimental evaluation with comparison to previous works.

Optimizing Access across Multiple Hierarchies in Data Warehouses

“…Xu et al [20] proposed a method for computing closest common subexpressions for view selection problems that works at the level of relational algebra and hence SQL statements. Our work differs in that we identify common sub-hierarchies among hierarchies.…”

Materialized Views Selection of Multi-Dimensional Data in Real-Time Active Data Warehouses

“…最小为原则,从给定的视图集合中选择一部分视图进行实化 [12] .代价模型是实视图的选择问题中的一个重要部 分,在计算代价时,有些方法只考虑查询代价,而有些方法则综合查询代价和视图维护代价.本文假设在计算代 价时,只考虑查询代价而不考虑视图维护代价.…”

Optimizing Hierarchical Access in OLAP Environment

“…Much research has been done on view selection and query rewriting to exploit views. Those previous work focus mainly on flat relational views [2], regular and balanced hierarchies [3], [4], [3], and cube views. The proposed techniques do no extend easily to the irregular hierarchies that are supported by the current state-of-the-art dedicated OLAP servers.…”