| |
| | Teradata Magazine Enterprise View: Tomorrow's real-time ETL imperative |
 | | Extract and load routines have to bring the data in as is, and then, depending on the business's needs, do the transformation either on the way into the data mart or warehouse (for OLAP and strategic analysis) or on a transaction-by-transaction basis inside of a data-mining algorithm. |
 | | While transformation mid-stream will never go away entirely, there is a growing need to tie transformation to in-database data mining, in-database data quality, in-database ranking and scoring, in-database integration, in-database parallelism, and in-database scalability, fault tolerance and fail-over. |
 | | For example, transformation elements such as data mining and data cleansing will be introduced downstream instead of upstream. |
| www.teradata.com /t/page/127085 (1852 words) |
|