发明名称 Discounted future value operations on a massively parallel processing system and methods thereof
摘要 A method for performing an iterative operation using columnar data on a massively parallel processing system includes arranging data from a data source in a columnar table format, receiving tabular data containing data points representing a forecast of future values, receiving instructions to perform an iterative calculation on the data points, and for a predetermined selection of the columnar data points, assigning portions of the predetermined selection to a parallel processing system and applying structured query language operators to the columnar data and the tabular data to calculate results of the iterative calculation. The method also includes analyzing the iterative operation to identify at least one invariant function, and creating a columnar data having data points representing the invariant function. A system for implementing the method and a non-transitory computer readable medium are also disclosed.
申请公布号 US9626397(B2) 申请公布日期 2017.04.18
申请号 US201313766318 申请日期 2013.02.13
申请人 Business Objects Software Limited 发明人 O'Carrol Luan
分类号 G06F17/30 主分类号 G06F17/30
代理机构 Buckley, Maschoff & Talwalkar LLC 代理人 Buckley, Maschoff & Talwalkar LLC
主权项 1. A computer-implemented method comprising: arranging, by a relational engine, data from a data source into columnar data points in a columnar table format; receiving one or more respective tabular data sets containing data points representing a discounted operation forecast of future values; receiving instructions to perform an iterative calculation on the data points; analyzing, by a database optimizer, the iterative calculation to identify at least one invariant function, the invariant function being a loop invariant function unaffected by input values; pre-calculating results of the at least one invariant function to create columnar data having data points representing the invariant function; and a parallel calculation engine assigning portions of the pre-calculated results to a parallel processing system and applying respective structured query language operator sets to the pre-calculated results and to respective multiple tabular data sets to calculate, in parallel, respective results of the iterative calculation.
地址 Dublin IE