Building the Data Warehouse

Скачать в pdf «Building the Data Warehouse»

If very high degrees of accuracy are desired, a useful technique is to formulate the request and go through the iterative processing on the living sample database. In doing so, the DSS analyst quickly formulates the request. Then, after several iterations of analysis have been done, when the request is understood, it is run one final time against the large database.

Living sample data is just one more way of changing the level of granularity in the data warehouse to accommodate DSS processing.

Partitioning as a Design Approach

A second major design issue of data in the warehouse (after that of granularity) is that of partitioning (see Figure 2.11b). Partitioning of data refers to the

breakup of data into separate physical units that can be handled independently. In the data warehouse, the issues surrounding partitioning do not focus on whether partitioning should be done but how it should be done.

It is often said that if both granularity and partitioning are done properly, then almost all other aspects of the data warehouse design and implementation come easily. If granularity is not handled properly and if partitioning is not designed and implemented carefully, then no other aspects of design really matter.

Proper partitioning can benefit the data warehouse in several ways:

■■ Loading data ■■ Accessing data ■■ Archiving data ■■ Deleting data ■■ Monitoring data

■    Storing data

Partitioning data properly allows data to grow and to be managed. Not partitioning data properly does not allow data to be managed or to grow gracefully.

Скачать в pdf «Building the Data Warehouse»