In ETL (extract, load, transform) operations, database extracts are fed into a staging database which performs the transformations.
Using native drivers and specialized parallel extraction techniques, IRI FACT (Fast Extract) rapidly unloads data from tables in: Oracle, DB2, Sybase, MySQL, SQL Server, Altibase, Greenplum, and Teradata.
FACT produces flat files containing the selected data with incredible speed. FACT uses SQL query syntax to help you get the data you need. Built-in formatting options support your target's formatting requirements.
By pre-sorting the extracts, you can dramatically improve the performance of any relational database load utility. Any of the high-performance sort utilities in IRI CoSort can pre-sort the flat files, along with web logs, COBOL index files, and other data sources bound for your data warehouse database or other applications.
In ELT, the database will transform. Use built-in ER diagramming, database profiling, and graphical SQL editing tools in the IRI Workbench IDE, built on Eclipse.
Consider also the chance to "'lighten the load" of transformations before you load, or "offload" transformations after you load. The SortCL program in IRI CoSort performs all the same data transformations that SQL does. But SortCL does them much faster, in a single I/O pass outside the database.
SortCL scripts are also much smaller and simpler, and use central metadata that is standard across all DB and flat-file sources. SortCL can transform the data in the flat-file extracts, or in the tables themselves.
Both FACT and CoSort are available separately, or as part of the IRI Data Manager software suite or IRI Voracity data management platform with visual workflow. You can design, run, and manage the operations of both products in the free IRI Workbench GUI.