Optimal chunking of large multidimensional arrays for data warehousing

ss domain. Using this more abstract approach means that more data sources of varying types can be incorporated with less effort, and such heterogeneous data sources might be very relevant in a future that can bring us the semantic web and perhaps other external sources. With regards to the processing of the ETL flow, which I have definitely done my share of on the physical level, I was impressed by the work of Vasiliki Tziovara, Panos Vassiliadis and Alkis Simitis titled "Deciding the Physical Implementation of ETL Workflows". This project proposes that we implement a declarative language for ETL in order to obtain detail independence and leave it to the system to implement the actual workflow, including the application of sorters. Again, from personal experiences, I have learned the hard way how important sorting and indexing can be during the implementation of a physical level ETL flow. This led me to suggest a "sort upon load" in the sense of applying a primary i...

• Type: Conference
• Year: 2007
• Where: DOLAP
• Authors: Ekow J. Otoo, Doron Rotem, Sridhar Seshadri
• Keywords: Data Sources, DOLAP 2007, Etl Flows, Information Management, Level Etl Flow