What: There sure is a lot of jargon in this field. Here' my summary of converting data from an operational database to a data warehouse. It goes something like this: Operational data are stored in OLTP software which consists of typical 2 dimensional relational tables. These work well for rapid and constant transactions: updates, insertions and deletions. To build a data warehouse, it's best to have data in a form in which it can be queried and displayed in a multitude of dimensions, which is the objective of having a data warehouse. Thus the conceptual model of the Rubiks cubes. But this is a conceptual model that helps humans visualize several dimensions, the data itself never "exists" in 3 dimensions. Data for a data warehouse are converted into and stored in a form which an OLAP tool can analyze it in a multitude of dimensions. This is the function of the star schema. It is a particular type of relational table structure in which 2 dimensional tables can be analyzed as if they were related to each other in a multitude of dimensions. A visualization crutch for this process might be the conversion that occurs with Origami, in which a 2-d object, a piece of paper, is converted into a 3-d object by systematic folding.

Writer: mattyflynn

LCC: QA

Where: Reference Link Has Evaporated

Date: May 15 2010 2:16 AM