Spreadsheets in RDBMS for OLAP

摘要: One of the critical deficiencies SQL is lack support for n-dimensional array-based computations which are frequent in OLAP environments. Relational (ROLAP) applications have to emulate them using joins, recently introduced Window Functions [18] and complex inefficient CASE expressions. The designated place specifying calculations SELECT clause, extremely limiting forces user generate queries nested views, subqueries joins. Furthermore, SQL-query optimizer pre-occupied with determining efficient join orders choosing optimal access methods largely disregards optimization numerical formulas. Execution concentrated on computation a cube [11], [16] rather than random structures inter-row calculations. This has created gap that been filled by spreadsheets specialized MOLAP engines, good at formulas mathematical modeling but formalism relational model, difficult manage, exhibit scalability problems. paper presents extensions involving array based modeling. In addition, we present optimizations, execution models processing efficiently.