Vertical Fragmentation of Views in Relational Data Warehouses

. Within the framework of the data warehouse design<E-345> methodology we are developing, in this paper we investigate the problem<E-338> of vertical fragmentation of relational views aimed at minimizing the<E-358> global query response time. Each view includes several measures which,<E-344> within the workload, are seldom requested together; thus, the system<E-343> performance may be increased by partitioning the views to be materialized<E-324> into smaller tables. On the other hand, drill-across queries involve measures<E-336> taken from two or more views; in this case the access costs may be decreased<E-335> by unifying these views into larger tables. Within the data warehouse<E-331> context, the presence of redundant views makes the fragmentation problem<E-338> more complex than in traditional relational databases since it requires to<E-362> decide on which views each query should be executed. After formalizing the<E-358> fragmentation problem as a 0-1 integer linear programming pr..